AI coding plans vs. API pricing: Claude, Copilot, Cursor, Z.ai and 20+ more compared

On this page

Measure the workload first#
What I would trial first#
Normalize the subscription costs#
Start with the API baseline#
Compare with OpenRouter too#
Convert the allowance before comparing it#
Native coding subscriptions belong in the trial too#
Temporary offers, kept separate#
The next analysis should measure accepted work#
Reproduce and update the comparison#
Changelog#
2026-09-25 — Every tracked subscription, first-party only#
2026-09-25 — First-party re-verification#
2026-09-07 — Published as a data artifact#
2026-09-07 — Display the analysis date#
2026-09-07 — Index by subscription and model#
2026-09-07 — OpenRouter route comparison#
2026-09-07 — Normalize subscription costs#
2026-09-07 — Measured usage, supported models and capacity#
2026-09-07 — Initial research draft#

agents September 7, 2026 updated September 25, 2026 112 min read

For terminals & agents: curl -s https://jedarden.com/data/inference-plans.md

Once coding workers can keep themselves busy, inference becomes a purchasing decision. A subscription buys an allowance that expires. An API bills for the work requested. Both can be economical, and both can be wasted.

In The unit economics of running cattle, I argued for measuring cost per completed outcome. This artifact applies that discipline to buying inference: which offers deserve a trial, what their advertised units mean, and where the apparent savings disappear.

Analysis date: September 25, 2026. Prices, quotas and routes were rechecked against the providers’ own pages that day; the usage measurement below is from September 7.

This is a maintained research artifact. Prices are in USD before tax; individual source checks and usage captures retain their own dates in the dataset. The calculations model billing; they do not establish model quality, sustained throughput, or the cost of a shipped application. The changelog records substantive revisions. Publication dates, editorial updates, and source-verification dates serve different purposes.

Measure the workload first#

The baseline now comes from an actual local usage scan captured September 7, 2026 at 14:55 UTC, using Tokscale 4.15.1. It covers the locally discovered Claude Code, Codex and ZCode histories in one scan, across 871,701 recorded messages. This is an accumulated history snapshot, not one month of usage or a deduplicated total across every machine. The aggregate measurement contains the exact counters and derivation; no prompts or session records are included.

Fresh input2,293,232,314

Share of all tokens: 2.794%
Per 1M output: 5.985M

Cache reads78,848,276,612

Share of all tokens: 96.082%
Per 1M output: 205.776M

Cache writes538,768,695

Share of all tokens: 0.657%
Per 1M output: 1.406M

Output, including billed reasoning383,174,954

Share of all tokens: 0.467%
Per 1M output: 1.000M

82.06B measured tokens; 213.2:1 total input/output; 96.53% of input served from cache. All input includes fresh input, cache reads and cache writes.

Tokscale’s model report keeps separately reported reasoning outside its output field. I add that bucket once to obtain billable output; reasoning already included by a client stays inside output. Cache reads and writes are separate from fresh input. This normalization follows the versioned scanner implementation.

These are measured usage proportions, transferred into hypothetical purchases below. Another model, tokenizer, provider cache or session pattern can change them. In particular, the high cache-hit rate is not guaranteed on another service. Most processed tokens here are repeated context; maximizing that total alone would reward repetition rather than useful applications.

What I would trial first#

My shortlist from the published economics is OpenCode Go for routine coding, Z.ai for sustained GLM-5.3 work, and Synthetic when its model selection or additional concurrency justifies a pack. Keep a capped API balance available for overflow and difficult tasks. These are candidates to measure in the actual harness, not a ranking of coding quality. The September 25 recheck keeps the shortlist but narrows Go’s role: OpenRouter’s cheapest GLM-5.3 route now undercuts Go’s GLM-5.3 allowance, so Go earns its place on MiMo, MiniMax and Flash-class models, and Z.ai remains the GLM-5.3 option.

Fifteen subscriptions were added on September 25, including the native coding products, China-region editions and several credit-based plans. Three of them publish enough to calculate and are worth measuring. StepFun’s Step Plan states the largest allowance per dollar here: its Credits convert at “approximately $1 ≈ 7M Credits” of list-price usage, with no five-hour or weekly window, so the $29 tier covers about 47B processed Step 3.5 Flash tokens at this mix. Its models have not been benchmarked here. Command Code sells $70 of list-price usage for $10 on GOAT, with per-model caps. GitHub Copilot joins annual Kilo Pass as a quantifiable way to buy Claude Sonnet 5 below Anthropic’s list price, and it also covers Opus 5.5, at 33–50% below list depending on tier. Its credits above the fee are a flex allotment that “may change over time”. ChatGPT, Claude, Cursor, Google AI, Devin and SuperGrok publish limits but no convertible allowance. Their rows quote those limits; turning them into token capacity would take a measurement of my own usage, which this revision does not claim.

The key eligibility question is whether the allowance covers the workload. A developer running a supported coding tool, an unattended coding worker, and a finished application serving customers can fall under different rules. An API-shaped endpoint does not establish permission for all three.

Each expanded plan reports total processed tokens, with output in parentheses, at the measured mix above. Each example spends the entire allowance on its named model; examples within one plan are alternatives. Weekly quotas are prorated to 30 days and assume full utilization within shorter limits. B means billion and M means million tokens. This is quota capacity, not a measured throughput promise or a model’s per-request context window.

OpenCode Go$1033 models / variants · 3 calculations

Supported models / catalog

MiMo V2.6 Flash
MiniMax M3
Kimi K3
DeepSeek V4 Flash
GLM-5.3
GPT-5.6 Luna
GPT 6 Luna
MiMo V2.6 Pro
MiMo V2.5
MiMo V2.5 Pro
GLM-5.3-Flash
DeepSeek V4 Pro
DeepSeek V4.1 Flash
Grok 4.6
Grok 4.7
GLM-5.2
GLM-5.1
Kimi K2.7 Code
Kimi K2.6
MiniMax M2.7
MiniMax M2.5
Qwen 3.8 Max
Qwen 3.8 Flash
Qwen 3.7 Max
Qwen 3.7 Plus
Qwen 3.6 Plus
LongCat 2.0
Muse Spark 1.3 Contributor
Muse Spark 1.2 Contributor
DeepSeek V4 Flash Vision Exp
Hy4 preview
Hy3
Space Bunny

Model source.

Estimated processed tokens / 30 days

MiMo V2.6 Flash6.80B total (31.73M output)
MiniMax M3815.16M total (3.81M output)
Kimi K332.48M total (0.15M output)

Allowance and practical limit

Monthly API allowance per model: $60 for MiMo V2.6 Flash, MiniMax M3, GLM-5.3-Flash, GLM-5.2/5.1, Kimi K2.7 Code/K2.6, Qwen3.7 Plus and LongCat 2.0; $30 for DeepSeek V4 Flash; $15 for Kimi K3, GLM-5.3, GPT 6 Luna and DeepSeek V4 Pro. One shared pool; each model is also capped at 20% per five hours and 50% per week. Coding-agent traffic only.

Z.ai Lite / Pro / Max$18 / $80 / $1682 models / variants · 3 calculations

Supported models / catalog

GLM-5.3
GLM-5.3-Flash

Model source.

Estimated processed tokens / 30 days

Lite, GLM-5.3 off-peak432.12M total (2.02M output)
Pro, GLM-5.3 off-peak2.59B total (12.11M output)
Max, GLM-5.3 off-peak6.05B total (28.25M output)

Allowance and practical limit

10k / 60k / 140k credits weekly; model and cache weights apply. Off-peak usage costs half the credits; peak is weekdays 14:00-18:00 UTC+8. Supported coding tools; dynamic concurrency.

Synthetic$30 per pack11 models / variants · 2 calculations

Supported models / catalog

syn:large:text
syn:small:text
syn:large:vision
syn:small:vision
openai/gpt-oss-120b
zai-org/GLM-5.3-Flash
deepseek-ai/DeepSeek-V4.1-Flash
moonshotai/Kimi-K3
Qwen/Qwen3.8-27B
zai-org/GLM-4.7-Flash
nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-NVFP4

Model source.

Estimated processed tokens / 30 days

GLM-5.3-Flash2.24B total (10.45M output)
Kimi-K3169.75M total (0.79M output)

Allowance and practical limit

$24 of its own API credits weekly; 500 weighted requests per five hours; one concurrent request per model per pack. More packs increase each limit.

camelStream$5 per stream4 models / variants · 1 calculation

Supported models / catalog

DeepSeek V4.1 Flash
GLM-5.3-Flash
GPT-5.6 Luna
Muse Spark 1.3

Model source.

Estimated processed tokens / 30 days

Unmetered; throughput unknown

Allowance and practical limit

Unmetered tokens; one generation per stream. Provider selects the model. Published throughput targets are not a service-level agreement; standard terms permit training on submitted content.

MiniMax Token Plan$22 / $55 / $1322 models / variants · 1 calculation

Supported models / catalog

MiniMax M3
MiniMax M2.7

Model source.

Estimated processed tokens / 30 days

Not quantifiable from published terms

Allowance and practical limit

Five-hour and weekly windows; text, image and speech share quota, deducted at pay-as-you-go list prices. The provider publishes conflicting token figures and none for current prices, so no API-equivalent is calculated.

Kimi membership / CodePlus $19 / Pro $39 / Max $99 / Ultra $1993 models / variants · 1 calculation

Supported models / catalog

Kimi K3
kimi-for-coding
kimi-for-coding-highspeed

Model source.

Estimated processed tokens / 30 days

Not quantifiable from published terms

Allowance and practical limit

Agent credits scale 1x / 2x / 5x / 10x by tier and share the membership monthly quota; the base amount is not published. New members have only a rolling five-hour window. Extra usage bills near Open Platform API prices. China-region CNY prices are served only to China-region sessions; the Chinese docs say pricing is unchanged under the new tier names.

Alibaba Cloud Coding Plan$5010 models / variants · 1 calculation

Supported models / catalog

qwen3.7-plus
qwen3.6-plus
kimi-k2.5
glm-5
MiniMax-M2.5
qwen3.5-plus
qwen3-max-2026-01-23
qwen3-coder-next
qwen3-coder-plus
glm-4.7

Model source.

Estimated processed tokens / 30 days

Not quantifiable from published terms

Allowance and practical limit

6k requests per five hours, 45k per week, 90k per month; a task can use 5-30+ model calls. Explicitly excludes automated scripts, application backends and noninteractive use. Limited slots, restocked daily.

deepseekv4pro.comCoding $19.90 / Coding Max $49.90 / Agent $49.90 / Coding Pro Max $1995 models / variants · 1 calculation

Supported models / catalog

DeepSeek V4.1 Flash
DeepSeek V4 Pro
GLM-5.3
MiniMax M3
Kimi K3

Model source.

Estimated processed tokens / 30 days

Not quantifiable from published terms

Allowance and practical limit

Independent reseller, not affiliated with DeepSeek. Recurring promotional prices and inventory-gated slots. Quotas are published in unweighted "plan tokens"; weighting and upstream plan rules apply.

MiMo Token Plan$6 / $16 / $50 / $1002 models / variants · 4 calculations

Supported models / catalog

MiMo V2.6 Flash
MiMo V2.6 Pro

Model source.

Estimated processed tokens / 30 days

Lite, V2.6 Flash daytime650.12M total (3.04M output)
Standard, V2.6 Flash daytime1.74B total (8.14M output)
Pro, V2.6 Flash daytime6.03B total (28.13M output)
Max, V2.6 Flash daytime13.00B total (60.71M output)

Allowance and practical limit

4.1B / 11B / 38B / 82B credits, with model-specific deductions. These are not raw tokens. Programming-tool restrictions apply.

Chutes$10 / $2014 models / variants · 2 calculations

Supported models / catalog

google/gemma-4-31B-turbo-TEE
Qwen/Qwen3.6-27B-TEE
Qwen/Qwen3.8-27B-TEE
Qwen/Qwen3.5-397B-A17B-TEE
zai-org/GLM-5.1-TEE
deepseek-ai/DeepSeek-V3.2-TEE
zai-org/GLM-5.2-TEE
moonshotai/Kimi-K2.6-TEE
deepseek-ai/DeepSeek-V4-Flash-0731-TEE
unsloth/Mistral-Nemo-Instruct-2407-TEE
moonshotai/Kimi-K3-TEE
Qwen/Qwen3-235B-A22B-Thinking-2507-TEE
Qwen/Qwen3-32B-TEE
Nemotron-3-Nano-Omni-30B-TEE

Model source.

Estimated processed tokens / 30 days

Plus, Flash conditional ceiling785.87M total (3.67M output)
Pro, Flash conditional ceiling1.57B total (7.34M output)

Allowance and practical limit

Daily quotas; overage is 6% (Plus) or 10% (Pro) below PAYG rates. The February policy caps subscription value at 5x its PAYG rates. Per-tier quotas are not published.

NanoGPT$12296 models / variants · 2 calculations

Supported models / catalog

GLM 5.3 Flash
MiniMax M3
MiMo V2.6 Flash
MiMo V2.6 Pro
GLM 5.3
DeepSeek V4 Pro 0813
DeepSeek V4.1 Flash

Showing 7 compared routes from 296 recorded models / variants. Browse the complete catalog.

Estimated processed tokens / 30 days

1× input weight258.35M total (1.21M output)
2× input weight129.17M total (0.60M output)

Allowance and practical limit

60M input-token units weekly, including cached input, with model multipliers; resets Monday 00:00 UTC. Shared or commercial team workloads must use pay-as-you-go.

Kilo Pass$19 / $49 / $199394 models / variants · 3 calculations

Supported models / catalog

DeepSeek: DeepSeek V4.1 Flash
Z.ai: GLM 5.3 Flash
MoonshotAI: Kimi K3
MiniMax: MiniMax M3
Xiaomi: MiMo-V2.6-Flash
Xiaomi: MiMo-V2.6-Pro
Z.ai: GLM 5.3
DeepSeek: DeepSeek V4 Pro 0813
Anthropic: Claude Sonnet 5

Showing 9 compared routes from 394 recorded models / variants. Browse the complete catalog.

Estimated processed tokens / 30 days

Starter, Sonnet 5 annual91.59M total (0.43M output)
Pro, Sonnet 5 annual236.21M total (1.10M output)
Expert, Sonnet 5 annual959.32M total (4.48M output)

Allowance and practical limit

Credits plus welcome/loyalty bonuses; annual billing gives 50% extra monthly credits. Full use yields a 33.3% effective discount. Bonus credits expire monthly; Gateway supported.

ChatGPT (Codex)Go $8 / Plus $20 / Pro from $1009 models / variants · 1 calculation

Supported models / catalog

GPT-6 Astra
GPT-6 Sol
GPT-6 Luna
GPT-5.6 Sol
GPT-5.6 Terra
GPT-5.6 Luna
GPT-5.5
GPT-5.4
GPT-5.4 mini

Model source.

Estimated processed tokens / 30 days

All tiers

Allowance and practical limit

Model-dependent five-hour message allowances, with weekly limits possible; Pro gives 5x or 20x Plus. Extra usage can be bought as ChatGPT credits. Per-token credit rates are published, but the included allowance is not.

Claude (Claude Code)Pro $20 / Max 5x $100 / Max 20x $2003 models / variants · 1 calculation

Supported models / catalog

Claude Fable 5.1
Claude Opus 5.5
Claude Sonnet 5

Model source.

Estimated processed tokens / 30 days

All tiers

Allowance and practical limit

Five-hour session limits plus a weekly limit on Max; Max is 5x or 20x Pro, whose allowance is not published. Limits assume ordinary, individual use; developers building products must use API keys.

CursorPro $20 / Pro+ $60 / Ultra $20013 models / variants · 1 calculation

Supported models / catalog

Grok 4.7
Grok 4.6
Grok 4.5
Composer 2.5
Claude Fable 5.1
Claude Opus 5.5
Claude Sonnet 5
Gemini 3.1 Pro
Gemini 3.8 Flash
GPT-5.6 Sol
GPT-5.6 Terra
GPT-5.6 Luna
Muse Spark 1.3

Model source.

Estimated processed tokens / 30 days

All tiers

Allowance and practical limit

Two monthly usage pools, one for Cursor's own models and one for third-party models at their API prices. Per-model rates are published; the included dollar amount per plan is not.

GitHub CopilotPro $10 / Pro+ $39 / Max $1003 models / variants · 3 calculations

Supported models / catalog

Claude Sonnet 5
Claude Opus 5.5
GPT-6 Luna

Model source.

Estimated processed tokens / 30 days

Pro, Claude Sonnet 548.21M total (0.23M output)
Pro+, Claude Sonnet 5224.97M total (1.05M output)
Max, Claude Sonnet 5642.76M total (3.00M output)

Allowance and practical limit

1,500 / 7,000 / 20,000 monthly AI credits (1 credit = $0.01), spent at per-token model prices. Base credits equal the fee; the rest is a flex allotment that "may change over time." Chat, CLI and cloud agent draw on it.

Google AI Pro / UltraPro $19.99 / Ultra $99.99 (5x) or $199.99 (20x)3 models / variants · 3 calculations

Supported models / catalog

Gemini CLI (Gemini model family)
Antigravity agent models
Jules

Model source.

Estimated processed tokens / 30 days

Pro: 1,500 CLI requests/day
Ultra 5x: 2,000 CLI requests/day
Ultra 20x: 2,000 CLI requests/day

Allowance and practical limit

Gemini CLI: 1,500 (Pro) or 2,000 (Ultra) model requests per user per day. Jules: 100 or 300 tasks per rolling 24 hours. Antigravity: tiered rate limits plus an AI credit pool; no token amounts published.

Devin Pro / MaxPro $20 / Max $2002 models / variants · 2 calculations

Supported models / catalog

SWE-2
Frontier and open-source models

Model source.

Estimated processed tokens / 30 days

Pro: SWE-2 free until Oct 10
Max: weekly quota, size unpublished

Allowance and practical limit

Pro has a daily and weekly usage quota; Max a larger weekly quota with no daily cap. Quota sizes are not published. Extra usage is prepaid on-demand credit at API pricing. SWE-2 is free in Devin Desktop and CLI through October 10, 2026.

SuperGrokSuperGrok $30 / Plus $100; Lite and Heavy prices not shown on the pricing page2 models / variants · 1 calculation

Supported models / catalog

Grok 4.6
Grok Build

Model source.

Estimated processed tokens / 30 days

Limits not quantified

Allowance and practical limit

Tiers: Lite, SuperGrok, Plus and Heavy. Limits are described only as "higher rate limits" and "significantly higher usage across Chat, Imagine, Voice & Build". No quantities are published.

Ollama Cloud Pro / MaxPro $20 ($200/yr) / Max $1006 models / variants · 3 calculations

Supported models / catalog

GLM-5.3
GLM-5.3-Flash
Kimi K3
MiniMax M3
DeepSeek V4.1 Flash
DeepSeek V4 Pro

Model source.

Estimated processed tokens / 30 days

Pro, GLM-5.3188.28M total (0.88M output)
Pro, DeepSeek V4.1 Flash off-peak5.52B total (25.80M output)
Max, GLM-5.3941.41M total (4.40M output)

Allowance and practical limit

Pro includes $60 and Max $300 of usage credits per month, deducted at the published per-token model prices; no rollover. Concurrency: Pro 3, Max 10 requests. Off-peak rates for DeepSeek apply outside weekday 12:00-18:00 UTC.

Command CodeGo $1 / GOAT $10 / Pro $20 / Max 10x $100 / Max 20x $2008 models / variants · 4 calculations

Supported models / catalog

DeepSeek V4 Flash
GLM-5.2
Kimi K2.7 Code
MiMo V2.6 Flash
Kimi K3
GLM-5.3
Claude Sonnet 5
Claude Opus 5.5

Model source.

Estimated processed tokens / 30 days

GOAT, GLM-5.2219.66M total (1.03M output)
GOAT, DeepSeek V4 Flash5.52B total (25.80M output)
Pro, Claude Sonnet 564.28M total (0.30M output)
Max 10x, Claude Opus 5.5232.47M total (1.09M output)

Allowance and practical limit

Monthly credit pools of $10 / $70 / $80 / $150 / $300 at Command Code's per-model token prices, with per-model caps on GOAT and Pro and separate standard/premium limits on Max. Every plan also caps 5-hour and weekly spend. Payment processing fee extra.

StepFun Step PlanFlash Mini $6.99 / Plus $9.99 / Pro $29 / Max $994 models / variants · 3 calculations

Supported models / catalog

Step 3.5 Flash
Step 5 Preview
Step 3.7 Flash
step-router-v1

Model source.

Estimated processed tokens / 30 days

Flash Mini, Step 3.5 Flash2.37B total (11.09M output)
Flash Plus, Step 3.5 Flash9.50B total (44.34M output)
Flash Max, Step 5 Preview60.05B total (280.39M output)

Allowance and practical limit

400M / 1,600M / 8,000M / 40,000M Credits per month, deducted at list price (about 7M Credits per $1). Monthly pool only, no five-hour or weekly window; booster packs available.

StepFun Step Plan (China)Flash Mini ¥49 / Plus ¥99 / Pro ¥199 / Max ¥6992 models / variants · 2 calculations

Supported models / catalog

Step 3.5 Flash
Step 5 Preview

Model source.

Estimated processed tokens / 30 days

Flash Mini, Step 3.5 Flash2.37B total (11.09M output)
Flash Max, Step 3.5 Flash237.42B total (1.11B output)

Allowance and practical limit

Same Credit pools as the international plan; 1M Credits = ¥1 of list-price usage. USD equivalents use 6.7489 CNY/USD (CFETS, 2026-09-24).

GLM Coding Plan (China)Lite ¥118 / Pro ¥538 / Max ¥1,0782 models / variants · 2 calculations

Supported models / catalog

GLM-5.3
GLM-5.3-Flash

Model source.

Estimated processed tokens / 30 days

Lite, GLM-5.3 off-peak432.12M total (2.02M output)
Max, GLM-5.3 off-peak6.05B total (28.25M output)

Allowance and practical limit

bigmodel.cn edition of the Z.ai plan: same 10k / 60k / 140k weekly credits and coefficients, priced in CNY (USD at 6.7489, CFETS 2026-09-24). Prices are rendered client-side; figures are the credits-based set in the official site code.

MiniMax Token Plan (China)Plus ¥49 / Max ¥119 / Ultra ¥4692 models / variants · 1 calculation

Supported models / catalog

MiniMax M3
MiniMax M2.7

Model source.

Estimated processed tokens / 30 days

All tiers

Allowance and practical limit

Five-hour and weekly windows; text, image and speech share one quota deducted at pay-as-you-go prices. No token allowance published except an approximate Ultra figure.

Alibaba Bailian Coding Plan (China)Pro ¥200 (first month ¥39.90)1 model / variant · 1 calculation

Supported models / catalog

qwen3.7-plus

Model source.

Estimated processed tokens / 30 days

Allowance and practical limit

CNY edition of the Alibaba Coding Plan: same request limits and model list. Explicitly excludes automated scripts, backends and noninteractive use; inputs are used for model improvement.

Alibaba Bailian Token Plan (China)Lite ¥39 / Essential ¥79 / Standard ¥139 / Pro ¥499 (limited-time; list ¥60 / ¥120 / ¥180 / ¥600)3 models / variants · 1 calculation

Supported models / catalog

qwen3.8-max
deepseek-v4.1-flash
glm-5.3

Model source.

Estimated processed tokens / 30 days

All tiers

Allowance and practical limit

Credit-based successor to the Coding Plan: monthly Credit caps, no weekly limit since 2026-09-22. Credits-per-token not published. Beijing region only; interactive tools only, no automation; one per verified identity.

Several qualifications matter more than another decimal place in the price:

Z.ai supports specific coding tools and uses dynamic concurrency limits. Its published examples include automated development tasks, but that does not establish unrestricted custom-backend access. Usage policy
camelStream’s standard terms permit retention and training on prompts and outputs, without an opt-out. Model choice, queue time and throughput are not guaranteed. It merits a small experiment with suitable non-sensitive work. Terms
Kimi Code’s personal benefits differ from production API access. The membership page alone does not establish a fixed Code token allowance or identical context limits across products. Code benefits
NanoGPT’s new subscription terms explicitly exclude building commercial products. They apply to new users now and existing users from September 14. PAYG permits commercial use. Terms, quota accounting
The DeepSeek reseller needs further clarification. Its pricing cards and DeepSeek integration page give different allowances. Its separate Agent Plan documentation describes an upstream Volcengine product. I have not treated either as an official DeepSeek subscription or assigned it a reliable savings multiplier.
Chutes’ old unlimited-value anecdotes are obsolete. Its published revision caps monthly subscription benefit at five times its own PAYG value and permits shorter rolling limits.

The capacity CSV includes additional models, peak/off-peak variants and all four token buckets. The catalog snapshot preserves the large public model lists. Model availability, subscription inclusion and a selectable API alias are distinct; unresolved tier or alias details are marked in the comparison.

Normalize the subscription costs#

The same measured mix makes every quantifiable allowance comparable in token units: 1M billable output tokens requires approximately 213.167M input tokens, for 214.167M total processed tokens. The input is divided among the fresh, cache-read and cache-write buckets above before applying each provider’s prices or credit weights. Dividing a subscription’s fee by its advertised credit count would skip that conversion.

output_M = allowance_for_period / deductions_per_1M_output_and_its_input
processed_M = output_M × (1 + measured_input_output_ratio)
cost_per_1M_processed = fee_for_period / processed_M
cost_per_1M_output_with_input = fee_for_period / output_M

These are fully utilized, steady-state 30-day estimates from the dated September 25 price snapshot. Output costs include the associated input and cache processing; they are not the provider’s output-only API rate. Processed-token costs include cache reads, which dominate this workload. Both measures describe the same workload and produce the same cost ordering. Differences in model quality still need an accepted-work benchmark.

OpenCode Go: GoMiMo V2.6 Flash$0.32 / 1M output

Fee / month: $10.00
Processed / 30 days: 6.80B
Output / 30 days: 31.73M
$ / 1M processed: $0.00147
$ / 1M output + its input: $0.32
Saving vs direct API: 83.3% less

OpenCode Go: GoMiniMax M3$2.63 / 1M output

Fee / month: $10.00
Processed / 30 days: 815.16M
Output / 30 days: 3.81M
$ / 1M processed: $0.01227
$ / 1M output + its input: $2.63
Saving vs direct API: 83.3% less

OpenCode Go: GoKimi K3$65.94 / 1M output

Fee / month: $10.00
Processed / 30 days: 32.48M
Output / 30 days: 0.15M
$ / 1M processed: $0.30788
$ / 1M output + its input: $65.94
Saving vs direct API: 33.3% less

Synthetic: One packGLM-5.3-Flash$2.87 / 1M output

Fee / month: $30.00
Processed / 30 days: 2.24B
Output / 30 days: 10.45M
$ / 1M processed: $0.01340
$ / 1M output + its input: $2.87
Saving vs direct API: 63.1% less

Synthetic: One packKimi-K3$37.85 / 1M output

Fee / month: $30.00
Processed / 30 days: 169.75M
Output / 30 days: 0.79M
$ / 1M processed: $0.17673
$ / 1M output + its input: $37.85
Saving vs direct API: 61.7% less

MiMo Token Plan: LiteMiMo V2.6 Flash daytime$1.98 / 1M output

Fee / month: $6.00
Processed / 30 days: 650.12M
Output / 30 days: 3.04M
$ / 1M processed: $0.00923
$ / 1M output + its input: $1.98
Saving vs direct API: 4.5% more expensive

MiMo Token Plan: StandardMiMo V2.6 Flash daytime$1.96 / 1M output

Fee / month: $16.00
Processed / 30 days: 1.74B
Output / 30 days: 8.14M
$ / 1M processed: $0.00917
$ / 1M output + its input: $1.96
Saving vs direct API: 3.9% more expensive

MiMo Token Plan: ProMiMo V2.6 Flash daytime$1.78 / 1M output

Fee / month: $50.00
Processed / 30 days: 6.03B
Output / 30 days: 28.13M
$ / 1M processed: $0.00830
$ / 1M output + its input: $1.78
Saving vs direct API: 6.0% less

MiMo Token Plan: MaxMiMo V2.6 Flash daytime$1.65 / 1M output

Fee / month: $100.00
Processed / 30 days: 13.00B
Output / 30 days: 60.71M
$ / 1M processed: $0.00769
$ / 1M output + its input: $1.65
Saving vs direct API: 12.9% less

Chutes: PlusDeepSeek V4 Flash · conditional ceiling$2.73 / 1M output

Fee / month: $10.00
Processed / 30 days: 785.87M
Output / 30 days: 3.67M
$ / 1M processed: $0.01272
$ / 1M output + its input: $2.73
Saving vs direct API: 17.2% more expensive

Chutes: ProDeepSeek V4 Flash · conditional ceiling$2.73 / 1M output

Fee / month: $20.00
Processed / 30 days: 1.57B
Output / 30 days: 7.34M
$ / 1M processed: $0.01272
$ / 1M output + its input: $2.73
Saving vs direct API: 17.2% more expensive

NanoGPT: ProIncluded model at 1x input weight$9.95 / 1M output

Fee / month: $12.00
Processed / 30 days: 258.35M
Output / 30 days: 1.21M
$ / 1M processed: $0.04645
$ / 1M output + its input: $9.95
Saving vs direct API: Not established

NanoGPT: ProIncluded model at 2x input weight$19.90 / 1M output

Fee / month: $12.00
Processed / 30 days: 129.17M
Output / 30 days: 0.60M
$ / 1M processed: $0.09290
$ / 1M output + its input: $19.90
Saving vs direct API: Not established

Kilo Pass: Starter annual billingAnthropic: Claude Sonnet 5$44.43 / 1M output

Fee / month: $19.00
Processed / 30 days: 91.59M
Output / 30 days: 0.43M
$ / 1M processed: $0.20744
$ / 1M output + its input: $44.43
Saving vs direct API: Not established

Kilo Pass: Pro annual billingAnthropic: Claude Sonnet 5$44.43 / 1M output

Fee / month: $49.00
Processed / 30 days: 236.21M
Output / 30 days: 1.10M
$ / 1M processed: $0.20744
$ / 1M output + its input: $44.43
Saving vs direct API: Not established

Kilo Pass: Expert annual billingAnthropic: Claude Sonnet 5$44.43 / 1M output

Fee / month: $199.00
Processed / 30 days: 959.32M
Output / 30 days: 4.48M
$ / 1M processed: $0.20744
$ / 1M output + its input: $44.43
Saving vs direct API: Not established

GitHub Copilot: ProClaude Sonnet 5$44.43 / 1M output

Fee / month: $10.00
Processed / 30 days: 48.21M
Output / 30 days: 0.23M
$ / 1M processed: $0.20744
$ / 1M output + its input: $44.43
Saving vs direct API: 33.3% less

GitHub Copilot: Pro+Claude Sonnet 5$37.13 / 1M output

Fee / month: $39.00
Processed / 30 days: 224.97M
Output / 30 days: 1.05M
$ / 1M processed: $0.17336
$ / 1M output + its input: $37.13
Saving vs direct API: 44.3% less

GitHub Copilot: MaxClaude Sonnet 5$33.32 / 1M output

Fee / month: $100.00
Processed / 30 days: 642.76M
Output / 30 days: 3.00M
$ / 1M processed: $0.15558
$ / 1M output + its input: $33.32
Saving vs direct API: 50.0% less

Ollama Cloud Pro / Max: ProGLM-5.3$22.75 / 1M output

Fee / month: $20.00
Processed / 30 days: 188.28M
Output / 30 days: 0.88M
$ / 1M processed: $0.10622
$ / 1M output + its input: $22.75
Saving vs direct API: 66.7% less

Ollama Cloud Pro / Max: MaxGLM-5.3$22.75 / 1M output

Fee / month: $100.00
Processed / 30 days: 941.41M
Output / 30 days: 4.40M
$ / 1M processed: $0.10622
$ / 1M output + its input: $22.75
Saving vs direct API: 66.7% less

Ollama Cloud Pro / Max: ProDeepSeek V4.1 Flash off-peak$0.78 / 1M output

Fee / month: $20.00
Processed / 30 days: 5.52B
Output / 30 days: 25.80M
$ / 1M processed: $0.00362
$ / 1M output + its input: $0.78
Saving vs direct API: 66.7% less

Command Code: GOATGLM-5.2$9.75 / 1M output

Fee / month: $10.00
Processed / 30 days: 219.66M
Output / 30 days: 1.03M
$ / 1M processed: $0.04552
$ / 1M output + its input: $9.75
Saving vs direct API: Not established

Command Code: GOATDeepSeek V4 Flash off-peak$0.39 / 1M output

Fee / month: $10.00
Processed / 30 days: 5.52B
Output / 30 days: 25.80M
$ / 1M processed: $0.00181
$ / 1M output + its input: $0.39
Saving vs direct API: 83.3% less

Command Code: ProClaude Sonnet 5$66.64 / 1M output

Fee / month: $20.00
Processed / 30 days: 64.28M
Output / 30 days: 0.30M
$ / 1M processed: $0.31116
$ / 1M output + its input: $66.64
Saving vs direct API: 0.0% less

Command Code: Max 10xClaude Opus 5.5$92.12 / 1M output

Fee / month: $100.00
Processed / 30 days: 232.47M
Output / 30 days: 1.09M
$ / 1M processed: $0.43015
$ / 1M output + its input: $92.12
Saving vs direct API: 0.0% more expensive

StepFun Step Plan: Flash MiniStep 3.5 Flash$0.63 / 1M output

Fee / month: $6.99
Processed / 30 days: 2.37B
Output / 30 days: 11.09M
$ / 1M processed: $0.00294
$ / 1M output + its input: $0.63
Saving vs direct API: 87.8% less

StepFun Step Plan (China): Flash MiniStep 3.5 Flash$0.65 / 1M output

Fee / month: $7.26
Processed / 30 days: 2.37B
Output / 30 days: 11.09M
$ / 1M processed: $0.00306
$ / 1M output + its input: $0.65
Saving vs direct API: Not established

StepFun Step Plan: Flash PlusStep 3.5 Flash$0.23 / 1M output

Fee / month: $9.99
Processed / 30 days: 9.50B
Output / 30 days: 44.34M
$ / 1M processed: $0.00105
$ / 1M output + its input: $0.23
Saving vs direct API: 95.6% less

StepFun Step Plan: Flash MaxStep 5 Preview$0.35 / 1M output

Fee / month: $99.00
Processed / 30 days: 60.05B
Output / 30 days: 280.39M
$ / 1M processed: $0.00165
$ / 1M output + its input: $0.35
Saving vs direct API: 98.3% less

StepFun Step Plan (China): Flash MaxStep 3.5 Flash$0.09 / 1M output

Fee / month: $103.57
Processed / 30 days: 237.42B
Output / 30 days: 1.11B
$ / 1M processed: $0.00044
$ / 1M output + its input: $0.09
Saving vs direct API: Not established

GLM Coding Plan (China): LiteGLM-5.3 (off-peak)$8.66 / 1M output

Fee / month: $17.48
Processed / 30 days: 432.12M
Output / 30 days: 2.02M
$ / 1M processed: $0.04045
$ / 1M output + its input: $8.66
Saving vs direct API: 88.3% less

GLM Coding Plan (China): MaxGLM-5.3 (off-peak)$5.65 / 1M output

Fee / month: $159.73
Processed / 30 days: 6.05B
Output / 30 days: 28.25M
$ / 1M processed: $0.02640
$ / 1M output + its input: $5.65
Saving vs direct API: 92.3% less

Z.ai Lite / Pro / Max: LiteGLM-5.3 (off-peak)$8.92 / 1M output

Fee / month: $18.00
Processed / 30 days: 432.12M
Output / 30 days: 2.02M
$ / 1M processed: $0.04166
$ / 1M output + its input: $8.92
Saving vs direct API: 86.9% less

Z.ai Lite / Pro / Max: ProGLM-5.3 (off-peak)$6.61 / 1M output

Fee / month: $80.00
Processed / 30 days: 2.59B
Output / 30 days: 12.11M
$ / 1M processed: $0.03086
$ / 1M output + its input: $6.61
Saving vs direct API: 90.3% less

Z.ai Lite / Pro / Max: MaxGLM-5.3 (off-peak)$5.95 / 1M output

Fee / month: $168.00
Processed / 30 days: 6.05B
Output / 30 days: 28.25M
$ / 1M processed: $0.02777
$ / 1M output + its input: $5.95
Saving vs direct API: 91.3% less

The comparison covers the quantifiable examples from the plan overview. Other model allocations and time windows use the same calculation:

Show the remaining model and time-of-day calculations

OpenCode Go: GoDeepSeek V4 Flash off-peak$0.78 / 1M output

Fee / month: $10.00
Processed / 30 days: 2.76B
Output / 30 days: 12.90M
$ / 1M processed: $0.00362
$ / 1M output + its input: $0.78
Saving vs direct API: 66.7% less

OpenCode Go: GoGLM-5.3$45.50 / 1M output

Fee / month: $10.00
Processed / 30 days: 47.07M
Output / 30 days: 0.22M
$ / 1M processed: $0.21245
$ / 1M output + its input: $45.50
Saving vs direct API: 33.3% less

OpenCode Go: GoGPT 6 Luna$2.22 / 1M output

Fee / month: $10.00
Processed / 30 days: 964.14M
Output / 30 days: 4.50M
$ / 1M processed: $0.01037
$ / 1M output + its input: $2.22
Saving vs direct API: Not established

OpenCode Go: GoGPT-5.6 Luna$4.58 / 1M output

Fee / month: $10.00
Processed / 30 days: 468.02M
Output / 30 days: 2.19M
$ / 1M processed: $0.02137
$ / 1M output + its input: $4.58
Saving vs direct API: Not established

Synthetic: One packgpt-oss-120b$1.45 / 1M output

Fee / month: $30.00
Processed / 30 days: 4.45B
Output / 30 days: 20.76M
$ / 1M processed: $0.00675
$ / 1M output + its input: $1.45
Saving vs direct API: Not established

Synthetic: One packDeepSeek-V4.1-Flash$3.44 / 1M output

Fee / month: $30.00
Processed / 30 days: 1.87B
Output / 30 days: 8.71M
$ / 1M processed: $0.01608
$ / 1M output + its input: $3.44
Saving vs direct API: 48.1% more expensive

Synthetic: One packQwen3.8-27B$7.01 / 1M output

Fee / month: $30.00
Processed / 30 days: 916.11M
Output / 30 days: 4.28M
$ / 1M processed: $0.03275
$ / 1M output + its input: $7.01
Saving vs direct API: Not established

Synthetic: One packGLM-4.7-Flash$1.56 / 1M output

Fee / month: $30.00
Processed / 30 days: 4.11B
Output / 30 days: 19.21M
$ / 1M processed: $0.00729
$ / 1M output + its input: $1.56
Saving vs direct API: Not established

Synthetic: One packNVIDIA-Nemotron-3-Super-120B-A12B-NVFP4$4.54 / 1M output

Fee / month: $30.00
Processed / 30 days: 1.42B
Output / 30 days: 6.61M
$ / 1M processed: $0.02120
$ / 1M output + its input: $4.54
Saving vs direct API: Not established

MiMo Token Plan: LiteMiMo V2.6 Flash off-peak$1.58 / 1M output

Fee / month: $6.00
Processed / 30 days: 812.66M
Output / 30 days: 3.79M
$ / 1M processed: $0.00738
$ / 1M output + its input: $1.58
Saving vs direct API: 16.4% less

MiMo Token Plan: LiteMiMo V2.6 Pro daytime$4.88 / 1M output

Fee / month: $6.00
Processed / 30 days: 263.55M
Output / 30 days: 1.23M
$ / 1M processed: $0.02277
$ / 1M output + its input: $4.88
Saving vs direct API: 1.0% more expensive

MiMo Token Plan: LiteMiMo V2.6 Pro off-peak$3.90 / 1M output

Fee / month: $6.00
Processed / 30 days: 329.44M
Output / 30 days: 1.54M
$ / 1M processed: $0.01821
$ / 1M output + its input: $3.90
Saving vs direct API: 19.2% less

MiMo Token Plan: StandardMiMo V2.6 Flash off-peak$1.57 / 1M output

Fee / month: $16.00
Processed / 30 days: 2.18B
Output / 30 days: 10.18M
$ / 1M processed: $0.00734
$ / 1M output + its input: $1.57
Saving vs direct API: 16.9% less

MiMo Token Plan: StandardMiMo V2.6 Pro daytime$4.85 / 1M output

Fee / month: $16.00
Processed / 30 days: 707.10M
Output / 30 days: 3.30M
$ / 1M processed: $0.02263
$ / 1M output + its input: $4.85
Saving vs direct API: 0.4% more expensive

MiMo Token Plan: StandardMiMo V2.6 Pro off-peak$3.88 / 1M output

Fee / month: $16.00
Processed / 30 days: 883.87M
Output / 30 days: 4.13M
$ / 1M processed: $0.01810
$ / 1M output + its input: $3.88
Saving vs direct API: 19.7% less

MiMo Token Plan: ProMiMo V2.6 Flash off-peak$1.42 / 1M output

Fee / month: $50.00
Processed / 30 days: 7.53B
Output / 30 days: 35.17M
$ / 1M processed: $0.00664
$ / 1M output + its input: $1.42
Saving vs direct API: 24.8% less

MiMo Token Plan: ProMiMo V2.6 Pro daytime$4.38 / 1M output

Fee / month: $50.00
Processed / 30 days: 2.44B
Output / 30 days: 11.41M
$ / 1M processed: $0.02047
$ / 1M output + its input: $4.38
Saving vs direct API: 9.2% less

MiMo Token Plan: ProMiMo V2.6 Pro off-peak$3.51 / 1M output

Fee / month: $50.00
Processed / 30 days: 3.05B
Output / 30 days: 14.26M
$ / 1M processed: $0.01638
$ / 1M output + its input: $3.51
Saving vs direct API: 27.3% less

MiMo Token Plan: MaxMiMo V2.6 Flash off-peak$1.32 / 1M output

Fee / month: $100.00
Processed / 30 days: 16.25B
Output / 30 days: 75.89M
$ / 1M processed: $0.00615
$ / 1M output + its input: $1.32
Saving vs direct API: 30.3% less

MiMo Token Plan: MaxMiMo V2.6 Pro daytime$4.06 / 1M output

Fee / month: $100.00
Processed / 30 days: 5.27B
Output / 30 days: 24.61M
$ / 1M processed: $0.01897
$ / 1M output + its input: $4.06
Saving vs direct API: 15.8% less

MiMo Token Plan: MaxMiMo V2.6 Pro off-peak$3.25 / 1M output

Fee / month: $100.00
Processed / 30 days: 6.59B
Output / 30 days: 30.77M
$ / 1M processed: $0.01518
$ / 1M output + its input: $3.25
Saving vs direct API: 32.6% less

Kilo Pass: Starter annual billingMiniMax: MiniMax M3$10.51 / 1M output

Fee / month: $19.00
Processed / 30 days: 387.20M
Output / 30 days: 1.81M
$ / 1M processed: $0.04907
$ / 1M output + its input: $10.51
Saving vs direct API: 33.3% less

Kilo Pass: Pro annual billingMiniMax: MiniMax M3$10.51 / 1M output

Fee / month: $49.00
Processed / 30 days: 998.57M
Output / 30 days: 4.66M
$ / 1M processed: $0.04907
$ / 1M output + its input: $10.51
Saving vs direct API: 33.3% less

Kilo Pass: Expert annual billingMiniMax: MiniMax M3$10.51 / 1M output

Fee / month: $199.00
Processed / 30 days: 4.06B
Output / 30 days: 18.94M
$ / 1M processed: $0.04907
$ / 1M output + its input: $10.51
Saving vs direct API: 33.3% less

GitHub Copilot: ProClaude Opus 5.5$61.42 / 1M output

Fee / month: $10.00
Processed / 30 days: 34.87M
Output / 30 days: 0.16M
$ / 1M processed: $0.28677
$ / 1M output + its input: $61.42
Saving vs direct API: 33.3% less

GitHub Copilot: ProGPT-6 Luna$2.22 / 1M output

Fee / month: $10.00
Processed / 30 days: 964.14M
Output / 30 days: 4.50M
$ / 1M processed: $0.01037
$ / 1M output + its input: $2.22
Saving vs direct API: 33.3% less

GitHub Copilot: Pro+Claude Opus 5.5$51.33 / 1M output

Fee / month: $39.00
Processed / 30 days: 162.73M
Output / 30 days: 0.76M
$ / 1M processed: $0.23966
$ / 1M output + its input: $51.33
Saving vs direct API: 44.3% less

GitHub Copilot: Pro+GPT-6 Luna$1.86 / 1M output

Fee / month: $39.00
Processed / 30 days: 4.50B
Output / 30 days: 21.01M
$ / 1M processed: $0.00867
$ / 1M output + its input: $1.86
Saving vs direct API: 44.3% less

GitHub Copilot: MaxClaude Opus 5.5$46.06 / 1M output

Fee / month: $100.00
Processed / 30 days: 464.95M
Output / 30 days: 2.17M
$ / 1M processed: $0.21508
$ / 1M output + its input: $46.06
Saving vs direct API: 50.0% less

GitHub Copilot: MaxGPT-6 Luna$1.67 / 1M output

Fee / month: $100.00
Processed / 30 days: 12.86B
Output / 30 days: 60.02M
$ / 1M processed: $0.00778
$ / 1M output + its input: $1.67
Saving vs direct API: 50.0% less

Ollama Cloud Pro / Max: ProGLM-5.3-Flash$2.59 / 1M output

Fee / month: $20.00
Processed / 30 days: 1.65B
Output / 30 days: 7.71M
$ / 1M processed: $0.01211
$ / 1M output + its input: $2.59
Saving vs direct API: 66.7% less

Ollama Cloud Pro / Max: MaxGLM-5.3-Flash$2.59 / 1M output

Fee / month: $100.00
Processed / 30 days: 8.26B
Output / 30 days: 38.55M
$ / 1M processed: $0.01211
$ / 1M output + its input: $2.59
Saving vs direct API: 66.7% less

Ollama Cloud Pro / Max: ProKimi K3$32.97 / 1M output

Fee / month: $20.00
Processed / 30 days: 129.92M
Output / 30 days: 0.61M
$ / 1M processed: $0.15394
$ / 1M output + its input: $32.97
Saving vs direct API: 66.7% less

Ollama Cloud Pro / Max: MaxKimi K3$32.97 / 1M output

Fee / month: $100.00
Processed / 30 days: 649.61M
Output / 30 days: 3.03M
$ / 1M processed: $0.15394
$ / 1M output + its input: $32.97
Saving vs direct API: 66.7% less

Ollama Cloud Pro / Max: ProMiniMax M3$10.51 / 1M output

Fee / month: $20.00
Processed / 30 days: 407.58M
Output / 30 days: 1.90M
$ / 1M processed: $0.04907
$ / 1M output + its input: $10.51
Saving vs direct API: 33.3% less

Ollama Cloud Pro / Max: MaxMiniMax M3$10.51 / 1M output

Fee / month: $100.00
Processed / 30 days: 2.04B
Output / 30 days: 9.52M
$ / 1M processed: $0.04907
$ / 1M output + its input: $10.51
Saving vs direct API: 33.3% less

Ollama Cloud Pro / Max: MaxDeepSeek V4.1 Flash off-peak$0.78 / 1M output

Fee / month: $100.00
Processed / 30 days: 27.62B
Output / 30 days: 128.98M
$ / 1M processed: $0.00362
$ / 1M output + its input: $0.78
Saving vs direct API: 66.7% less

Ollama Cloud Pro / Max: ProDeepSeek V4 Pro off-peak$3.80 / 1M output

Fee / month: $20.00
Processed / 30 days: 1.13B
Output / 30 days: 5.27M
$ / 1M processed: $0.01772
$ / 1M output + its input: $3.80
Saving vs direct API: 66.7% less

Ollama Cloud Pro / Max: MaxDeepSeek V4 Pro off-peak$3.80 / 1M output

Fee / month: $100.00
Processed / 30 days: 5.64B
Output / 30 days: 26.35M
$ / 1M processed: $0.01772
$ / 1M output + its input: $3.80
Saving vs direct API: 66.7% less

Command Code: GoDeepSeek V4 Flash off-peak$0.23 / 1M output

Fee / month: $1.00
Processed / 30 days: 920.77M
Output / 30 days: 4.30M
$ / 1M processed: $0.00109
$ / 1M output + its input: $0.23
Saving vs direct API: 90.0% less

Command Code: GOATKimi K2.7 Code$8.35 / 1M output

Fee / month: $10.00
Processed / 30 days: 256.39M
Output / 30 days: 1.20M
$ / 1M processed: $0.03900
$ / 1M output + its input: $8.35
Saving vs direct API: Not established

Command Code: GOATMiMo V2.6 Flash$0.95 / 1M output

Fee / month: $10.00
Processed / 30 days: 2.27B
Output / 30 days: 10.58M
$ / 1M processed: $0.00441
$ / 1M output + its input: $0.95
Saving vs direct API: 50.0% less

Command Code: GOATKimi K3$49.45 / 1M output

Fee / month: $10.00
Processed / 30 days: 43.31M
Output / 30 days: 0.20M
$ / 1M processed: $0.23091
$ / 1M output + its input: $49.45
Saving vs direct API: 50.0% less

Command Code: GOATGLM-5.3$34.12 / 1M output

Fee / month: $10.00
Processed / 30 days: 62.76M
Output / 30 days: 0.29M
$ / 1M processed: $0.15934
$ / 1M output + its input: $34.12
Saving vs direct API: 50.0% less

Command Code: ProDeepSeek V4 Flash off-peak$0.66 / 1M output

Fee / month: $20.00
Processed / 30 days: 6.45B
Output / 30 days: 30.10M
$ / 1M processed: $0.00310
$ / 1M output + its input: $0.66
Saving vs direct API: 71.4% less

Command Code: Max 10xDeepSeek V4 Flash off-peak$1.55 / 1M output

Fee / month: $100.00
Processed / 30 days: 13.81B
Output / 30 days: 64.49M
$ / 1M processed: $0.00724
$ / 1M output + its input: $1.55
Saving vs direct API: 33.3% less

Command Code: Max 20xDeepSeek V4 Flash off-peak$1.55 / 1M output

Fee / month: $200.00
Processed / 30 days: 27.62B
Output / 30 days: 128.98M
$ / 1M processed: $0.00724
$ / 1M output + its input: $1.55
Saving vs direct API: 33.3% less

Command Code: Max 20xClaude Opus 5.5$92.12 / 1M output

Fee / month: $200.00
Processed / 30 days: 464.95M
Output / 30 days: 2.17M
$ / 1M processed: $0.43015
$ / 1M output + its input: $92.12
Saving vs direct API: 0.0% more expensive

StepFun Step Plan: Flash MiniStep 5 Preview$2.49 / 1M output

Fee / month: $6.99
Processed / 30 days: 600.51M
Output / 30 days: 2.80M
$ / 1M processed: $0.01164
$ / 1M output + its input: $2.49
Saving vs direct API: 87.8% less

StepFun Step Plan: Flash PlusStep 5 Preview$0.89 / 1M output

Fee / month: $9.99
Processed / 30 days: 2.40B
Output / 30 days: 11.22M
$ / 1M processed: $0.00416
$ / 1M output + its input: $0.89
Saving vs direct API: 95.6% less

StepFun Step Plan (China): Flash PlusStep 3.5 Flash$0.33 / 1M output

Fee / month: $14.67
Processed / 30 days: 9.50B
Output / 30 days: 44.34M
$ / 1M processed: $0.00154
$ / 1M output + its input: $0.33
Saving vs direct API: Not established

StepFun Step Plan: Flash ProStep 3.5 Flash$0.13 / 1M output

Fee / month: $29.00
Processed / 30 days: 47.48B
Output / 30 days: 221.72M
$ / 1M processed: $0.00061
$ / 1M output + its input: $0.13
Saving vs direct API: 97.5% less

StepFun Step Plan: Flash ProStep 5 Preview$0.52 / 1M output

Fee / month: $29.00
Processed / 30 days: 12.01B
Output / 30 days: 56.08M
$ / 1M processed: $0.00241
$ / 1M output + its input: $0.52
Saving vs direct API: 97.5% less

StepFun Step Plan (China): Flash ProStep 3.5 Flash$0.13 / 1M output

Fee / month: $29.49
Processed / 30 days: 47.48B
Output / 30 days: 221.72M
$ / 1M processed: $0.00062
$ / 1M output + its input: $0.13
Saving vs direct API: Not established

StepFun Step Plan: Flash MaxStep 3.5 Flash$0.09 / 1M output

Fee / month: $99.00
Processed / 30 days: 237.42B
Output / 30 days: 1.11B
$ / 1M processed: $0.00042
$ / 1M output + its input: $0.09
Saving vs direct API: 98.3% less

GLM Coding Plan (China): LiteGLM-5.3 (peak)$17.33 / 1M output

Fee / month: $17.48
Processed / 30 days: 216.06M
Output / 30 days: 1.01M
$ / 1M processed: $0.08090
$ / 1M output + its input: $17.33
Saving vs direct API: 76.6% less

GLM Coding Plan (China): ProGLM-5.3 (off-peak)$6.59 / 1M output

Fee / month: $79.72
Processed / 30 days: 2.59B
Output / 30 days: 12.11M
$ / 1M processed: $0.03075
$ / 1M output + its input: $6.59
Saving vs direct API: 91.1% less

GLM Coding Plan (China): ProGLM-5.3 (peak)$13.17 / 1M output

Fee / month: $79.72
Processed / 30 days: 1.30B
Output / 30 days: 6.05M
$ / 1M processed: $0.06150
$ / 1M output + its input: $13.17
Saving vs direct API: 82.2% less

GLM Coding Plan (China): MaxGLM-5.3 (peak)$11.31 / 1M output

Fee / month: $159.73
Processed / 30 days: 3.02B
Output / 30 days: 14.12M
$ / 1M processed: $0.05281
$ / 1M output + its input: $11.31
Saving vs direct API: 84.7% less

Z.ai Lite / Pro / Max: LiteGLM-5.3 (peak)$17.84 / 1M output

Fee / month: $18.00
Processed / 30 days: 216.06M
Output / 30 days: 1.01M
$ / 1M processed: $0.08331
$ / 1M output + its input: $17.84
Saving vs direct API: 73.9% less

Z.ai Lite / Pro / Max: LiteGLM-5.3-Flash list (off-peak)$2.94 / 1M output

Fee / month: $18.00
Processed / 30 days: 1.31B
Output / 30 days: 6.11M
$ / 1M processed: $0.01375
$ / 1M output + its input: $2.94
Saving vs direct API: 62.2% less

Z.ai Lite / Pro / Max: LiteGLM-5.3-Flash list (peak)$5.89 / 1M output

Fee / month: $18.00
Processed / 30 days: 654.52M
Output / 30 days: 3.06M
$ / 1M processed: $0.02750
$ / 1M output + its input: $5.89
Saving vs direct API: 24.3% less

Z.ai Lite / Pro / Max: ProGLM-5.3 (peak)$13.22 / 1M output

Fee / month: $80.00
Processed / 30 days: 1.30B
Output / 30 days: 6.05M
$ / 1M processed: $0.06171
$ / 1M output + its input: $13.22
Saving vs direct API: 80.6% less

Z.ai Lite / Pro / Max: ProGLM-5.3-Flash list (off-peak)$2.18 / 1M output

Fee / month: $80.00
Processed / 30 days: 7.85B
Output / 30 days: 36.67M
$ / 1M processed: $0.01019
$ / 1M output + its input: $2.18
Saving vs direct API: 72.0% less

Z.ai Lite / Pro / Max: ProGLM-5.3-Flash list (peak)$4.36 / 1M output

Fee / month: $80.00
Processed / 30 days: 3.93B
Output / 30 days: 18.34M
$ / 1M processed: $0.02037
$ / 1M output + its input: $4.36
Saving vs direct API: 43.9% less

Z.ai Lite / Pro / Max: MaxGLM-5.3 (peak)$11.89 / 1M output

Fee / month: $168.00
Processed / 30 days: 3.02B
Output / 30 days: 14.12M
$ / 1M processed: $0.05554
$ / 1M output + its input: $11.89
Saving vs direct API: 82.6% less

Z.ai Lite / Pro / Max: MaxGLM-5.3-Flash list (off-peak)$1.96 / 1M output

Fee / month: $168.00
Processed / 30 days: 18.33B
Output / 30 days: 85.57M
$ / 1M processed: $0.00917
$ / 1M output + its input: $1.96
Saving vs direct API: 74.8% less

Z.ai Lite / Pro / Max: MaxGLM-5.3-Flash list (peak)$3.93 / 1M output

Fee / month: $168.00
Processed / 30 days: 9.16B
Output / 30 days: 42.79M
$ / 1M processed: $0.01833
$ / 1M output + its input: $3.93
Saving vs direct API: 49.5% less

“Saving vs direct API” compares the same named model against the direct API row below, using ordinary list prices for GLM Flash and off-peak prices for DeepSeek. Serving configurations and cache behavior can differ. “Not established” means this dataset lacks a defensible same-model baseline; it does not mean zero savings. Chutes estimates remain conditional on its published ceiling and model eligibility. Kilo examples assume annual billing with the full monthly bonus consumed before expiry. Alternative model allocations within a subscription cannot be added together.

The capacity CSV contains all four token buckets, both normalized costs, direct API savings and break-even utilization. It also separates a discount against the serving provider’s own PAYG rates from a discount against the original model API. Missing or conflicting quotas remain blank; camelStream’s unmetered service has no defensible fixed token capacity or normalized cost here. MiniMax, Kimi and the DeepSeek reseller need quota clarification; Alibaba needs tokens per billable request.

At 50% utilization, each subscription’s effective cost per token doubles. Break-even utilization is the fraction of modeled allowance that must replace direct API spending to recover the fee. A figure above 100% means even full utilization is more expensive than that API baseline. Real concurrency, request and reset limits can prevent reaching the modeled allowance.

Start with the API baseline#

The priced workload is 1M billable output tokens plus the measured proportions of fresh input, cache reads and cache writes shown above. Tokenizers differ, so equal token counts do not mean identical amounts of source code. These are estimates at current prices, not the historical cash cost of the measured sessions.

For prices stated per million tokens:

cost = fresh_input_M × input_price
     + cache_read_M × cache_read_price
     + cache_write_M × cache_creation_price
     + output_M × output_price
     + additional_charges

The comparison excludes paid tools, search, additional retries, infrastructure and long-context premiums. Cache creation uses the ordinary input price unless the serving provider specifies a separate full creation rate. Free cache storage or a waived write surcharge does not make the initial input processing free. Kilo’s Sonnet 5 example, for instance, uses its explicit cache-creation rate in the capacity calculator.

MiMo V2.6 Flash$1.89 / 1M output

Input / M: $0.14
Cached / M: $0.0028
Output / M: $0.28
Cost per 1M output + its input: $1.89
$100 processed capacity: 11.33B total (52.88M output)

DeepSeek V4.1 Flash off-peak$2.33 / 1M output

Input / M: $0.15
Cached / M: $0.003
Output / M: $0.60
Cost per 1M output + its input: $2.33
$100 processed capacity: 9.21B total (42.99M output)

GPT-6 Luna$3.33 / 1M output

Input / M: $0.1
Cached / M: $0.01
Output / M: $0.50
Cost per 1M output + its input: $3.33
$100 processed capacity: 6.43B total (30.01M output)

MiMo V2.6 Pro$4.83 / 1M output

Input / M: $0.435
Cached / M: $0.0036
Output / M: $0.87
Cost per 1M output + its input: $4.83
$100 processed capacity: 4.44B total (20.72M output)

Step 3.5 Flash$5.15 / 1M output

Input / M: $0.1
Cached / M: $0.02
Output / M: $0.30
Cost per 1M output + its input: $5.15
$100 processed capacity: 4.15B total (19.40M output)

GLM-5.3-Flash list$7.78 / 1M output

Input / M: $0.15
Cached / M: $0.03
Output / M: $0.50
Cost per 1M output + its input: $7.78
$100 processed capacity: 2.75B total (12.85M output)

DeepSeek V4 Pro off-peak$11.39 / 1M output

Input / M: $0.66
Cached / M: $0.022
Output / M: $1.98
Cost per 1M output + its input: $11.39
$100 processed capacity: 1.88B total (8.78M output)

MiniMax M3$15.76 / 1M output

Input / M: $0.3
Cached / M: $0.06
Output / M: $1.20
Cost per 1M output + its input: $15.76
$100 processed capacity: 1.36B total (6.34M output)

Step 5 Preview$20.38 / 1M output

Input / M: $1
Cached / M: $0.05
Output / M: $2.70
Cost per 1M output + its input: $20.38
$100 processed capacity: 1.05B total (4.91M output)

Claude Sonnet 5$66.64 / 1M output

Input / M: $2
Cached / M: $0.2
Output / M: $10.00
Cost per 1M output + its input: $66.64
$100 processed capacity: 321.38M total (1.50M output)

GLM-5.3$68.25 / 1M output

Input / M: $1.4
Cached / M: $0.26
Output / M: $4.40
Cost per 1M output + its input: $68.25
$100 processed capacity: 313.80M total (1.47M output)

GLM-5.3 (bigmodel.cn)$73.89 / 1M output

Input / M: $1.18538
Cached / M: $0.29634
Output / M: $4.15
Cost per 1M output + its input: $73.89
$100 processed capacity: 289.85M total (1.35M output)

Claude Opus 5.5$92.12 / 1M output

Input / M: $4
Cached / M: $0.2
Output / M: $20.00
Cost per 1M output + its input: $92.12
$100 processed capacity: 232.47M total (1.09M output)

Kimi K3$98.91 / 1M output

Input / M: $3
Cached / M: $0.3
Output / M: $15.00
Cost per 1M output + its input: $98.91
$100 processed capacity: 216.54M total (1.01M output)

Without cache hits, the same workload costs $30.12 on MiMo V2.6 Flash, $32.58 on off-peak DeepSeek V4.1 Flash, and $654.50 on Kimi K3. A plan that counts cached input at full token weight can lose much of its advantage against an API with inexpensive cache reads. Repeated context also inflates token totals without representing newly generated work.

DeepSeek’s listed prices apply outside weekday peak windows of 01:00–04:00 and 06:00–10:00 UTC, and all day on Chinese public holidays; peak rates double. DeepSeek has retired its V4 Flash model names and now serves them with V4.1 Flash at the V4.1 price. MiniMax M3’s listed rate applies through 512K input per request. These timing and context conditions belong in the comparison, not in a footnote to the invoice. DeepSeek pricing, MiniMax pricing

Compare with OpenRouter too#

OpenRouter narrows some subscription discounts because its alternative providers can undercut the model author’s API. The result depends heavily on cached-input pricing. The following snapshot was captured on September 25, 2026, at 13:49 UTC from OpenRouter’s public model-endpoint API. Each quote keeps one provider’s input, cache and output rates together.

The comparison is grouped by subscription, then model, then tier or time window. All twenty-seven researched coding subscriptions appear, including those with unknown costs. Models from the saved catalogs remain visible in the index; large unpriced catalogs expand underneath their subscription. Catalog presence, subscription eligibility and model pinning are separate questions, so each group retains those conditions. Native workflow subscriptions are discussed separately below.

The unit-cost fields report cash cost per 1M output tokens plus 213.167M associated input tokens, using the measured cache mix. Subscription summaries show the monthly fee, and capacity details show total processed tokens with output in parentheses. Each calculated scenario spends the full allowance on its named model; scenarios within one subscription are alternative allocations. “Not established” means the snapshot lacks a defensible entitlement or calculation; “Not quoted” means no matching OpenRouter route was priced.

OpenRouter figures include its standard card-purchase fee: 5.5%, with a $0.80 minimum. The reference purchase is $100 of inference credits plus $5.50 in fees, so quoted inference cost is multiplied by 1.055 once. A $10 credit purchase instead incurs an 8% effective fee. Taxes and special billing arrangements are excluded. OpenRouter fee policy

Jump to subscription: OpenCode Go · Z.ai Lite / Pro / Max · Synthetic · camelStream · MiniMax Token Plan · Kimi membership / Code · Alibaba Cloud Coding Plan · deepseekv4pro.com · MiMo Token Plan · Chutes · NanoGPT · Kilo Pass · ChatGPT (Codex) · Claude (Claude Code) · Cursor · GitHub Copilot · Google AI Pro / Ultra · Devin Pro / Max · SuperGrok · Ollama Cloud Pro / Max · Command Code · StepFun Step Plan · StepFun Step Plan (China) · GLM Coding Plan (China) · MiniMax Token Plan (China) · Alibaba Bailian Coding Plan (China) · Alibaba Bailian Token Plan (China)

OpenCode Go33 recorded models / aliases$10 / month

One shared allowance. Regional and experimental conditions apply; uncalculated models retain unknown capacity.

MiMo V2.6 Flash$0.32 subscription

Tier / conditions: Go
Processed / 30 days (output): 6.80B (31.73M output)
Subscription $ / 1M output + input: $0.32
OpenRouter $ / 1M output + input: $1.99Xiaomi, FP8
Subscription saving: 84.2% less

MiniMax M3$2.63 subscription

Tier / conditions: Go
Processed / 30 days (output): 815.16M (3.81M output)
Subscription $ / 1M output + input: $2.63
OpenRouter $ / 1M output + input: $13.30GMICloud, FP8
Subscription saving: 80.3% less

Kimi K3$65.94 subscription

Tier / conditions: Go
Processed / 30 days (output): 32.48M (0.15M output)
Subscription $ / 1M output + input: $65.94
OpenRouter $ / 1M output + input: $73.04Relace, FP4
Subscription saving: 9.7% less

DeepSeek V4 Flash$0.78 subscription

Tier / conditions: Go; off-peak
Processed / 30 days (output): 2.76B (12.90M output)
Subscription $ / 1M output + input: $0.78
OpenRouter $ / 1M output + input: $1.23Morph
Subscription saving: 36.8% less

GLM-5.3$45.50 subscription

Tier / conditions: Go
Processed / 30 days (output): 47.07M (0.22M output)
Subscription $ / 1M output + input: $45.50
OpenRouter $ / 1M output + input: $27.28Morph
Subscription saving: 66.8% more expensive

GPT-5.6 Luna$4.58 subscription

Tier / conditions: Go
Processed / 30 days (output): 468.02M (2.19M output)
Subscription $ / 1M output + input: $4.58
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

GPT 6 Luna$2.22 subscription

Tier / conditions: Go
Processed / 30 days (output): 964.14M (4.50M output)
Subscription $ / 1M output + input: $2.22
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

MiMo V2.6 Pro$5.09 OpenRouter

Tier / conditions: Go
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: $5.09Xiaomi, FP8
Subscription saving: Not established

MiMo V2.5Pricing not established

Tier / conditions: Go; deprecated 2026-10-21
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

MiMo V2.5 ProPricing not established

Tier / conditions: Go; deprecated 2026-10-21
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

GLM-5.3-Flash$2.67 OpenRouter

Tier / conditions: Go
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: $2.67InferenceNet, FP4
Subscription saving: Not established

DeepSeek V4 Pro$6.25 OpenRouter

Tier / conditions: Go
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: $6.25Baidu, FP8
Subscription saving: Not established

DeepSeek V4.1 Flash$1.23 OpenRouter

Tier / conditions: Go
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: $1.23Morph
Subscription saving: Not established

Grok 4.6Pricing not established

Tier / conditions: Go
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Grok 4.7Pricing not established

Tier / conditions: Go
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

GLM-5.2Pricing not established

Tier / conditions: Go
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

GLM-5.1Pricing not established

Tier / conditions: Go
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Kimi K2.7 CodePricing not established

Tier / conditions: Go
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Kimi K2.6Pricing not established

Tier / conditions: Go
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

MiniMax M2.7Pricing not established

Tier / conditions: Go
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

MiniMax M2.5Pricing not established

Tier / conditions: Go
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Qwen 3.8 MaxPricing not established

Tier / conditions: Go
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Qwen 3.8 FlashPricing not established

Tier / conditions: Go
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Qwen 3.7 MaxPricing not established

Tier / conditions: Go
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Qwen 3.7 PlusPricing not established

Tier / conditions: Go
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Qwen 3.6 PlusPricing not established

Tier / conditions: Go
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

LongCat 2.0Pricing not established

Tier / conditions: Go
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Muse Spark 1.3 ContributorPricing not established

Tier / conditions: Go
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Muse Spark 1.2 ContributorPricing not established

Tier / conditions: Go
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

DeepSeek V4 Flash Vision ExpPricing not established

Tier / conditions: Go
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Hy4 previewPricing not established

Tier / conditions: Go
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Hy3Pricing not established

Tier / conditions: Go
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Space BunnyPricing not established

Tier / conditions: Go; free for a limited time
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Z.ai Lite / Pro / Max2 recorded models / aliases$18 / $80 / $168 / month

Older aliases route to these models. Each row allocates the entire tier allowance to one model and time window.

GLM-5.36 scenarios

Tier / conditions: Lite; off-peak
Processed / 30 days (output): 432.12M (2.02M output)
Subscription $ / 1M output + input: $8.92
OpenRouter $ / 1M output + input: $27.28Morph
Subscription saving: 67.3% less

Tier / conditions: Lite; peak
Processed / 30 days (output): 216.06M (1.01M output)
Subscription $ / 1M output + input: $17.84
OpenRouter $ / 1M output + input: $27.28Morph
Subscription saving: 34.6% less

Tier / conditions: Pro; off-peak
Processed / 30 days (output): 2.59B (12.11M output)
Subscription $ / 1M output + input: $6.61
OpenRouter $ / 1M output + input: $27.28Morph
Subscription saving: 75.8% less

Tier / conditions: Pro; peak
Processed / 30 days (output): 1.30B (6.05M output)
Subscription $ / 1M output + input: $13.22
OpenRouter $ / 1M output + input: $27.28Morph
Subscription saving: 51.5% less

Tier / conditions: Max; off-peak
Processed / 30 days (output): 6.05B (28.25M output)
Subscription $ / 1M output + input: $5.95
OpenRouter $ / 1M output + input: $27.28Morph
Subscription saving: 78.2% less

Tier / conditions: Max; peak
Processed / 30 days (output): 3.02B (14.12M output)
Subscription $ / 1M output + input: $11.89
OpenRouter $ / 1M output + input: $27.28Morph
Subscription saving: 56.4% less

GLM-5.3-Flash6 scenarios

Tier / conditions: Lite; off-peak
Processed / 30 days (output): 1.31B (6.11M output)
Subscription $ / 1M output + input: $2.94
OpenRouter $ / 1M output + input: $2.67InferenceNet, FP4
Subscription saving: 10.3% more expensive

Tier / conditions: Lite; peak
Processed / 30 days (output): 654.52M (3.06M output)
Subscription $ / 1M output + input: $5.89
OpenRouter $ / 1M output + input: $2.67InferenceNet, FP4
Subscription saving: 120.6% more expensive

Tier / conditions: Pro; off-peak
Processed / 30 days (output): 7.85B (36.67M output)
Subscription $ / 1M output + input: $2.18
OpenRouter $ / 1M output + input: $2.67InferenceNet, FP4
Subscription saving: 18.3% less

Tier / conditions: Pro; peak
Processed / 30 days (output): 3.93B (18.34M output)
Subscription $ / 1M output + input: $4.36
OpenRouter $ / 1M output + input: $2.67InferenceNet, FP4
Subscription saving: 63.4% more expensive

Tier / conditions: Max; off-peak
Processed / 30 days (output): 18.33B (85.57M output)
Subscription $ / 1M output + input: $1.96
OpenRouter $ / 1M output + input: $2.67InferenceNet, FP4
Subscription saving: 26.5% less

Tier / conditions: Max; peak
Processed / 30 days (output): 9.16B (42.79M output)
Subscription $ / 1M output + input: $3.93
OpenRouter $ / 1M output + input: $2.67InferenceNet, FP4
Subscription saving: 47.1% more expensive

Synthetic11 recorded models / aliases$30 per pack / month

The saved served catalog includes four automatic aliases. Catalog availability alone does not establish subscription eligibility.

syn:large:textPricing not established

Tier / conditions: See plan terms
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

syn:small:textPricing not established

Tier / conditions: See plan terms
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

syn:large:visionPricing not established

Tier / conditions: See plan terms
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

syn:small:visionPricing not established

Tier / conditions: See plan terms
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

openai/gpt-oss-120b$1.45 subscription

Tier / conditions: One pack
Processed / 30 days (output): 4.45B (20.76M output)
Subscription $ / 1M output + input: $1.45
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

zai-org/GLM-5.3-Flash$2.87 subscription

Tier / conditions: One pack
Processed / 30 days (output): 2.24B (10.45M output)
Subscription $ / 1M output + input: $2.87
OpenRouter $ / 1M output + input: $2.67InferenceNet, FP4
Subscription saving: 7.5% more expensive

deepseek-ai/DeepSeek-V4.1-Flash$3.44 subscription

Tier / conditions: One pack
Processed / 30 days (output): 1.87B (8.71M output)
Subscription $ / 1M output + input: $3.44
OpenRouter $ / 1M output + input: $1.23Morph
Subscription saving: 180.7% more expensive

moonshotai/Kimi-K3$37.85 subscription

Tier / conditions: One pack
Processed / 30 days (output): 169.75M (0.79M output)
Subscription $ / 1M output + input: $37.85
OpenRouter $ / 1M output + input: $73.04Relace, FP4
Subscription saving: 48.2% less

Qwen/Qwen3.8-27B$7.01 subscription

Tier / conditions: One pack
Processed / 30 days (output): 916.11M (4.28M output)
Subscription $ / 1M output + input: $7.01
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

zai-org/GLM-4.7-Flash$1.56 subscription

Tier / conditions: One pack
Processed / 30 days (output): 4.11B (19.21M output)
Subscription $ / 1M output + input: $1.56
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-NVFP4$4.54 subscription

Tier / conditions: One pack
Processed / 30 days (output): 1.42B (6.61M output)
Subscription $ / 1M output + input: $4.54
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

camelStream4 recorded models / aliases$5 per stream / month

Unmetered service has no fixed token entitlement. Named routes are possible automatic choices, not selectable allocations.

DeepSeek V4.1 Flash$1.23 OpenRouter

Tier / conditions: One stream; automatic routing, no model pinning
Processed / 30 days (output): Unmetered; throughput unknown
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: $1.23Morph
Subscription saving: Not established

GLM-5.3-Flash$2.67 OpenRouter

Tier / conditions: One stream; automatic routing, no model pinning
Processed / 30 days (output): Unmetered; throughput unknown
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: $2.67InferenceNet, FP4
Subscription saving: Not established

GPT-5.6 LunaPricing not established

Tier / conditions: One stream; automatic routing, no model pinning
Processed / 30 days (output): Unmetered; throughput unknown
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Muse Spark 1.3Pricing not established

Tier / conditions: One stream; automatic routing, no model pinning
Processed / 30 days (output): Unmetered; throughput unknown
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

MiniMax Token Plan2 recorded models / aliases$22 / $55 / $132 / month

Text models only. Image and speech share quota but are outside this token comparison; published numerical allowance is insufficient.

MiniMax M3$13.30 OpenRouter

Tier / conditions: All listed tiers
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: $13.30GMICloud, FP8
Subscription saving: Not established

MiniMax M2.7Pricing not established

Tier / conditions: All listed tiers
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Kimi membership / Code3 recorded models / aliasesPlus $19 / Pro $39 / Max $99 / Ultra $199 / month

No fixed comparable entitlement established. Membership K3 and Code aliases are distinguished; alias costs are not assumed to equal K3.

Kimi K3$73.04 OpenRouter

Tier / conditions: Plus and above; 1M context on Pro and above
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: $73.04Relace, FP4
Subscription saving: Not established

kimi-for-codingPricing not established

Tier / conditions: Code alias; currently K2.8 Preview
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

kimi-for-coding-highspeedPricing not established

Tier / conditions: Code HighSpeed alias; currently K2.7 Code; Pro and above
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Alibaba Cloud Coding Plan10 recorded models / aliases$50 / month

Request quotas cannot be converted without tokens per billable request. Automated scripts and application backends are excluded.

qwen3.7-plusPricing not established

Tier / conditions: All listed tiers
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

qwen3.6-plusPricing not established

Tier / conditions: All listed tiers
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

kimi-k2.5Pricing not established

Tier / conditions: All listed tiers
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

glm-5Pricing not established

Tier / conditions: All listed tiers
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

MiniMax-M2.5Pricing not established

Tier / conditions: All listed tiers
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

qwen3.5-plusPricing not established

Tier / conditions: All listed tiers
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

qwen3-max-2026-01-23Pricing not established

Tier / conditions: All listed tiers
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

qwen3-coder-nextPricing not established

Tier / conditions: All listed tiers
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

qwen3-coder-plusPricing not established

Tier / conditions: All listed tiers
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

glm-4.7Pricing not established

Tier / conditions: All listed tiers
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

deepseekv4pro.com5 recorded models / aliasesCoding $19.90 / Coding Max $49.90 / Agent $49.90 / Coding Pro Max $199 / month

Independent reseller; published quotas conflict. Models added by the Agent plan are marked separately.

DeepSeek V4.1 Flash$1.23 OpenRouter

Tier / conditions: Coding tiers
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: $1.23Morph
Subscription saving: Not established

DeepSeek V4 Pro$6.25 OpenRouter

Tier / conditions: Agent plan (as DeepSeek V4)
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: $6.25Baidu, FP8
Subscription saving: Not established

GLM-5.3$27.28 OpenRouter

Tier / conditions: Agent plan only
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: $27.28Morph
Subscription saving: Not established

MiniMax M3$13.30 OpenRouter

Tier / conditions: Agent plan only
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: $13.30GMICloud, FP8
Subscription saving: Not established

Kimi K3$73.04 OpenRouter

Tier / conditions: Agent plan only
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: $73.04Relace, FP4
Subscription saving: Not established

MiMo Token Plan2 recorded models / aliases$6 / $16 / $50 / $100 / month

Text models only. Credit pools are shared across models; daytime and off-peak rows are alternative allocations.

MiMo V2.6 Flash8 scenarios

Tier / conditions: Lite; daytime
Processed / 30 days (output): 650.12M (3.04M output)
Subscription $ / 1M output + input: $1.98
OpenRouter $ / 1M output + input: $1.99Xiaomi, FP8
Subscription saving: 0.9% less

Tier / conditions: Lite; off-peak
Processed / 30 days (output): 812.66M (3.79M output)
Subscription $ / 1M output + input: $1.58
OpenRouter $ / 1M output + input: $1.99Xiaomi, FP8
Subscription saving: 20.7% less

Tier / conditions: Standard; daytime
Processed / 30 days (output): 1.74B (8.14M output)
Subscription $ / 1M output + input: $1.96
OpenRouter $ / 1M output + input: $1.99Xiaomi, FP8
Subscription saving: 1.5% less

Tier / conditions: Standard; off-peak
Processed / 30 days (output): 2.18B (10.18M output)
Subscription $ / 1M output + input: $1.57
OpenRouter $ / 1M output + input: $1.99Xiaomi, FP8
Subscription saving: 21.2% less

Tier / conditions: Pro; daytime
Processed / 30 days (output): 6.03B (28.13M output)
Subscription $ / 1M output + input: $1.78
OpenRouter $ / 1M output + input: $1.99Xiaomi, FP8
Subscription saving: 10.9% less

Tier / conditions: Pro; off-peak
Processed / 30 days (output): 7.53B (35.17M output)
Subscription $ / 1M output + input: $1.42
OpenRouter $ / 1M output + input: $1.99Xiaomi, FP8
Subscription saving: 28.7% less

Tier / conditions: Max; daytime
Processed / 30 days (output): 13.00B (60.71M output)
Subscription $ / 1M output + input: $1.65
OpenRouter $ / 1M output + input: $1.99Xiaomi, FP8
Subscription saving: 17.4% less

Tier / conditions: Max; off-peak
Processed / 30 days (output): 16.25B (75.89M output)
Subscription $ / 1M output + input: $1.32
OpenRouter $ / 1M output + input: $1.99Xiaomi, FP8
Subscription saving: 33.9% less

MiMo V2.6 Pro8 scenarios

Tier / conditions: Lite; daytime
Processed / 30 days (output): 263.55M (1.23M output)
Subscription $ / 1M output + input: $4.88
OpenRouter $ / 1M output + input: $5.09Xiaomi, FP8
Subscription saving: 4.2% less

Tier / conditions: Lite; off-peak
Processed / 30 days (output): 329.44M (1.54M output)
Subscription $ / 1M output + input: $3.90
OpenRouter $ / 1M output + input: $5.09Xiaomi, FP8
Subscription saving: 23.4% less

Tier / conditions: Standard; daytime
Processed / 30 days (output): 707.10M (3.30M output)
Subscription $ / 1M output + input: $4.85
OpenRouter $ / 1M output + input: $5.09Xiaomi, FP8
Subscription saving: 4.8% less

Tier / conditions: Standard; off-peak
Processed / 30 days (output): 883.87M (4.13M output)
Subscription $ / 1M output + input: $3.88
OpenRouter $ / 1M output + input: $5.09Xiaomi, FP8
Subscription saving: 23.9% less

Tier / conditions: Pro; daytime
Processed / 30 days (output): 2.44B (11.41M output)
Subscription $ / 1M output + input: $4.38
OpenRouter $ / 1M output + input: $5.09Xiaomi, FP8
Subscription saving: 13.9% less

Tier / conditions: Pro; off-peak
Processed / 30 days (output): 3.05B (14.26M output)
Subscription $ / 1M output + input: $3.51
OpenRouter $ / 1M output + input: $5.09Xiaomi, FP8
Subscription saving: 31.1% less

Tier / conditions: Max; daytime
Processed / 30 days (output): 5.27B (24.61M output)
Subscription $ / 1M output + input: $4.06
OpenRouter $ / 1M output + input: $5.09Xiaomi, FP8
Subscription saving: 20.2% less

Tier / conditions: Max; off-peak
Processed / 30 days (output): 6.59B (30.77M output)
Subscription $ / 1M output + input: $3.25
OpenRouter $ / 1M output + input: $5.09Xiaomi, FP8
Subscription saving: 36.2% less

Chutes14 recorded models / aliases$10 / $20 / month

Served catalog; exact subscription tier eligibility needs account confirmation. Flash capacity remains a conditional ceiling.

google/gemma-4-31B-turbo-TEEPricing not established

Tier / conditions: See plan terms
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Qwen/Qwen3.6-27B-TEEPricing not established

Tier / conditions: See plan terms
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Qwen/Qwen3.8-27B-TEEPricing not established

Tier / conditions: See plan terms
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Qwen/Qwen3.5-397B-A17B-TEEPricing not established

Tier / conditions: See plan terms
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

zai-org/GLM-5.1-TEEPricing not established

Tier / conditions: See plan terms
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

deepseek-ai/DeepSeek-V3.2-TEEPricing not established

Tier / conditions: See plan terms
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

zai-org/GLM-5.2-TEEPricing not established

Tier / conditions: See plan terms
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

moonshotai/Kimi-K2.6-TEEPricing not established

Tier / conditions: See plan terms
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

deepseek-ai/DeepSeek-V4-Flash-0731-TEE2 scenarios

Tier / conditions: Pinned V4 Flash 0731 checkpoint; compared with the V4.1 Flash routes DeepSeek now serves at the same list price; conditional ceiling
Processed / 30 days (output): 785.87M (3.67M output)
Subscription $ / 1M output + input: $2.73
OpenRouter $ / 1M output + input: $1.23Morph
Subscription saving: 122.1% more expensive

Tier / conditions: Pinned V4 Flash 0731 checkpoint; compared with the V4.1 Flash routes DeepSeek now serves at the same list price; conditional ceiling
Processed / 30 days (output): 1.57B (7.34M output)
Subscription $ / 1M output + input: $2.73
OpenRouter $ / 1M output + input: $1.23Morph
Subscription saving: 122.1% more expensive

unsloth/Mistral-Nemo-Instruct-2407-TEEPricing not established

Tier / conditions: See plan terms
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

moonshotai/Kimi-K3-TEE$73.04 OpenRouter

Tier / conditions: See plan terms
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: $73.04Relace, FP4
Subscription saving: Not established

Qwen/Qwen3-235B-A22B-Thinking-2507-TEEPricing not established

Tier / conditions: See plan terms
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Qwen/Qwen3-32B-TEEPricing not established

Tier / conditions: See plan terms
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Nemotron-3-Nano-Omni-30B-TEEPricing not established

Tier / conditions: See plan terms
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

NanoGPT296 recorded models / aliases$12 / month

Subscription-included model/variant catalog. Per-model multipliers were not captured, so the 1x/2x quota examples cannot be assigned to individual IDs.

Unassigned quota examples

Included model at 1x input weight$9.95 / 1M output + input
Included model at 2x input weight$19.90 / 1M output + input

Per-model weights remain unknown.

GLM 5.3 Flash$2.67 OpenRouter

Tier / conditions: See plan terms
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: $2.67InferenceNet, FP4
Subscription saving: Not established

MiniMax M3$13.30 OpenRouter

Tier / conditions: See plan terms
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: $13.30GMICloud, FP8
Subscription saving: Not established

MiMo V2.6 Flash$1.99 OpenRouter

Tier / conditions: See plan terms
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: $1.99Xiaomi, FP8
Subscription saving: Not established

MiMo V2.6 Pro$5.09 OpenRouter

Tier / conditions: See plan terms
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: $5.09Xiaomi, FP8
Subscription saving: Not established

GLM 5.3$27.28 OpenRouter

Tier / conditions: See plan terms
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: $27.28Morph
Subscription saving: Not established

DeepSeek V4 Pro 0813$6.25 OpenRouter

Tier / conditions: See plan terms
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: $6.25Baidu, FP8
Subscription saving: Not established

DeepSeek V4.1 Flash$1.23 OpenRouter

Tier / conditions: See plan terms
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: $1.23Morph
Subscription saving: Not established

Show 289 additional NanoGPT catalog models / variants

These IDs have no matched allowance or OpenRouter quote in this snapshot. Catalog coverage: Subscription-included endpoint.

Kilo Pass394 recorded models / aliases$19 / $49 / $199 / month

Gateway catalog includes free and paid variants. Annual credit bonus is modeled only where provider rates were captured; catalog presence alone does not establish eligibility.

DeepSeek: DeepSeek V4.1 Flash$1.23 OpenRouter

Tier / conditions: See plan terms
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: $1.23Morph
Subscription saving: Not established

Z.ai: GLM 5.3 Flash$2.67 OpenRouter

Tier / conditions: See plan terms
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: $2.67InferenceNet, FP4
Subscription saving: Not established

MoonshotAI: Kimi K3$73.04 OpenRouter

Tier / conditions: See plan terms
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: $73.04Relace, FP4
Subscription saving: Not established

MiniMax: MiniMax M33 scenarios

Tier / conditions: Starter annual billing
Processed / 30 days (output): 387.20M (1.81M output)
Subscription $ / 1M output + input: $10.51
OpenRouter $ / 1M output + input: $13.30GMICloud, FP8
Subscription saving: 21.0% less

Tier / conditions: Pro annual billing
Processed / 30 days (output): 998.57M (4.66M output)
Subscription $ / 1M output + input: $10.51
OpenRouter $ / 1M output + input: $13.30GMICloud, FP8
Subscription saving: 21.0% less

Tier / conditions: Expert annual billing
Processed / 30 days (output): 4.06B (18.94M output)
Subscription $ / 1M output + input: $10.51
OpenRouter $ / 1M output + input: $13.30GMICloud, FP8
Subscription saving: 21.0% less

Xiaomi: MiMo-V2.6-Flash$1.99 OpenRouter

Tier / conditions: See plan terms
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: $1.99Xiaomi, FP8
Subscription saving: Not established

Xiaomi: MiMo-V2.6-Pro$5.09 OpenRouter

Tier / conditions: See plan terms
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: $5.09Xiaomi, FP8
Subscription saving: Not established

Z.ai: GLM 5.3$27.28 OpenRouter

Tier / conditions: See plan terms
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: $27.28Morph
Subscription saving: Not established

DeepSeek: DeepSeek V4 Pro 0813$6.25 OpenRouter

Tier / conditions: See plan terms
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: $6.25Baidu, FP8
Subscription saving: Not established

Anthropic: Claude Sonnet 53 scenarios

Tier / conditions: Starter annual billing
Processed / 30 days (output): 91.59M (0.43M output)
Subscription $ / 1M output + input: $44.43
OpenRouter $ / 1M output + input: $70.31Anthropic
Subscription saving: 36.8% less

Tier / conditions: Pro annual billing
Processed / 30 days (output): 236.21M (1.10M output)
Subscription $ / 1M output + input: $44.43
OpenRouter $ / 1M output + input: $70.31Anthropic
Subscription saving: 36.8% less

Tier / conditions: Expert annual billing
Processed / 30 days (output): 959.32M (4.48M output)
Subscription $ / 1M output + input: $44.43
OpenRouter $ / 1M output + input: $70.31Anthropic
Subscription saving: 36.8% less

Show 385 additional Kilo Pass catalog models / variants

These IDs have no matched allowance or OpenRouter quote in this snapshot. Catalog coverage: Served catalog; consult plan terms for eligibility.

ChatGPT (Codex)9 recorded models / aliasesGo $8 / Plus $20 / Pro from $100 / month

Message estimates per five hours are published per model and tier; they are ranges, not a convertible token allowance. The Pro 20x price is not stated on the Codex pricing page and chatgpt.com/pricing blocks automated retrieval.

GPT-6 AstraPricing not established

Tier / conditions: All listed tiers
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

GPT-6 SolPricing not established

Tier / conditions: All listed tiers
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

GPT-6 LunaPricing not established

Tier / conditions: All listed tiers
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

GPT-5.6 SolPricing not established

Tier / conditions: All listed tiers
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

GPT-5.6 TerraPricing not established

Tier / conditions: All listed tiers
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

GPT-5.6 LunaPricing not established

Tier / conditions: All listed tiers
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

GPT-5.5Pricing not established

Tier / conditions: All listed tiers
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

GPT-5.4Pricing not established

Tier / conditions: All listed tiers
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

GPT-5.4 miniPricing not established

Tier / conditions: All listed tiers
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Claude (Claude Code)3 recorded models / aliasesPro $20 / Max 5x $100 / Max 20x $200 / month

Tier multiples are published; the base allowance is not, so no token capacity is calculated.

Claude Fable 5.1Pricing not established

Tier / conditions: All listed tiers
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Claude Opus 5.5Pricing not established

Tier / conditions: All listed tiers
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Claude Sonnet 5$70.31 OpenRouter

Tier / conditions: All listed tiers
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: $70.31Anthropic
Subscription saving: Not established

Cursor13 recorded models / aliasesPro $20 / Pro+ $60 / Ultra $200 / month

Per-model rates are published, but the included amount per plan is not, so no token capacity is calculated.

Grok 4.7Pricing not established

Tier / conditions: Cursor Models pool
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Grok 4.6Pricing not established

Tier / conditions: Cursor Models pool
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Grok 4.5Pricing not established

Tier / conditions: Cursor Models pool
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Composer 2.5Pricing not established

Tier / conditions: Cursor Models pool
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Claude Fable 5.1Pricing not established

Tier / conditions: Other Models pool, at API price
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Claude Opus 5.5Pricing not established

Tier / conditions: Other Models pool, at API price
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Claude Sonnet 5$70.31 OpenRouter

Tier / conditions: Other Models pool, at API price
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: $70.31Anthropic
Subscription saving: Not established

Gemini 3.1 ProPricing not established

Tier / conditions: Other Models pool, at API price
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Gemini 3.8 FlashPricing not established

Tier / conditions: Other Models pool, at API price
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

GPT-5.6 SolPricing not established

Tier / conditions: Other Models pool, at API price
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

GPT-5.6 TerraPricing not established

Tier / conditions: Other Models pool, at API price
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

GPT-5.6 LunaPricing not established

Tier / conditions: Other Models pool, at API price
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Muse Spark 1.3Pricing not established

Tier / conditions: Other Models pool, at API price
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

GitHub Copilot3 recorded models / aliasesPro $10 / Pro+ $39 / Max $100 / month

Only three models are calculated; every listed model draws on the same credit pool at its own rates. Paid plans get 10% off model costs under auto model selection, which is excluded here.

Claude Sonnet 53 scenarios

Tier / conditions: Pro
Processed / 30 days (output): 48.21M (0.23M output)
Subscription $ / 1M output + input: $44.43
OpenRouter $ / 1M output + input: $70.31Anthropic
Subscription saving: 36.8% less

Tier / conditions: Pro+
Processed / 30 days (output): 224.97M (1.05M output)
Subscription $ / 1M output + input: $37.13
OpenRouter $ / 1M output + input: $70.31Anthropic
Subscription saving: 47.2% less

Tier / conditions: Max
Processed / 30 days (output): 642.76M (3.00M output)
Subscription $ / 1M output + input: $33.32
OpenRouter $ / 1M output + input: $70.31Anthropic
Subscription saving: 52.6% less

Claude Opus 5.53 scenarios

Tier / conditions: Pro
Processed / 30 days (output): 34.87M (0.16M output)
Subscription $ / 1M output + input: $61.42
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Tier / conditions: Pro+
Processed / 30 days (output): 162.73M (0.76M output)
Subscription $ / 1M output + input: $51.33
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Tier / conditions: Max
Processed / 30 days (output): 464.95M (2.17M output)
Subscription $ / 1M output + input: $46.06
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

GPT-6 Luna3 scenarios

Tier / conditions: Pro
Processed / 30 days (output): 964.14M (4.50M output)
Subscription $ / 1M output + input: $2.22
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Tier / conditions: Pro+
Processed / 30 days (output): 4.50B (21.01M output)
Subscription $ / 1M output + input: $1.86
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Tier / conditions: Max
Processed / 30 days (output): 12.86B (60.02M output)
Subscription $ / 1M output + input: $1.67
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Google AI Pro / Ultra3 recorded models / aliasesPro $19.99 / Ultra $99.99 (5x) or $199.99 (20x) / month

Request, task and rate-limit allowances only. No tokens-per-request figure is published, so no token capacity is calculated.

Gemini CLI (Gemini model family)3 scenarios

Tier / conditions: Requests per user per day
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Tier / conditions: Requests per user per day
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Tier / conditions: Requests per user per day
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Antigravity agent modelsPricing not established

Tier / conditions: Pro: more generous rate limits; Ultra 5x / 20x: higher and highest rate limits
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

JulesPricing not established

Tier / conditions: Pro 100 / Ultra 300 tasks per rolling 24 hours; 15 / 60 concurrent
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Devin Pro / Max2 recorded models / aliasesPro $20 / Max $200 / month

Quota amounts are not published; the on-demand credit rates are also not published as per-token prices.

SWE-22 scenarios

Tier / conditions: Free in Devin Desktop and CLI through 2026-10-10
Processed / 30 days (output): Unmetered; throughput unknown
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Tier / conditions: Free in Devin Desktop and CLI through 2026-10-10
Processed / 30 days (output): Unmetered; throughput unknown
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Frontier and open-source models2 scenarios

Tier / conditions: Draws on the plan quota, then on-demand credits
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Tier / conditions: Draws on the plan quota, then on-demand credits
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

SuperGrok2 recorded models / aliasesSuperGrok $30 / Plus $100; Lite and Heavy prices not shown on the pricing page / month

No numerical usage limits are published for any tier.

Grok 4.6Pricing not established

Tier / conditions: Paid tiers
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Grok BuildPricing not established

Tier / conditions: All plans; usage scales by tier
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Ollama Cloud Pro / Max6 recorded models / aliasesPro $20 ($200/yr) / Max $100 / month

Dollar credits at Ollama's own rates; MiniMax M3 is priced at twice MiniMax's list rate.

GLM-5.32 scenarios

Tier / conditions: Pro and Max
Processed / 30 days (output): 188.28M (0.88M output)
Subscription $ / 1M output + input: $22.75
OpenRouter $ / 1M output + input: $27.28Morph
Subscription saving: 16.6% less

Tier / conditions: Pro and Max
Processed / 30 days (output): 941.41M (4.40M output)
Subscription $ / 1M output + input: $22.75
OpenRouter $ / 1M output + input: $27.28Morph
Subscription saving: 16.6% less

GLM-5.3-Flash2 scenarios

Tier / conditions: Pro and Max
Processed / 30 days (output): 1.65B (7.71M output)
Subscription $ / 1M output + input: $2.59
OpenRouter $ / 1M output + input: $2.67InferenceNet, FP4
Subscription saving: 2.8% less

Tier / conditions: Pro and Max
Processed / 30 days (output): 8.26B (38.55M output)
Subscription $ / 1M output + input: $2.59
OpenRouter $ / 1M output + input: $2.67InferenceNet, FP4
Subscription saving: 2.8% less

Kimi K32 scenarios

Tier / conditions: Pro and Max
Processed / 30 days (output): 129.92M (0.61M output)
Subscription $ / 1M output + input: $32.97
OpenRouter $ / 1M output + input: $73.04Relace, FP4
Subscription saving: 54.9% less

Tier / conditions: Pro and Max
Processed / 30 days (output): 649.61M (3.03M output)
Subscription $ / 1M output + input: $32.97
OpenRouter $ / 1M output + input: $73.04Relace, FP4
Subscription saving: 54.9% less

MiniMax M32 scenarios

Tier / conditions: Pro and Max
Processed / 30 days (output): 407.58M (1.90M output)
Subscription $ / 1M output + input: $10.51
OpenRouter $ / 1M output + input: $13.30GMICloud, FP8
Subscription saving: 21.0% less

Tier / conditions: Pro and Max
Processed / 30 days (output): 2.04B (9.52M output)
Subscription $ / 1M output + input: $10.51
OpenRouter $ / 1M output + input: $13.30GMICloud, FP8
Subscription saving: 21.0% less

DeepSeek V4.1 Flash2 scenarios

Tier / conditions: Off-peak rate; off-peak
Processed / 30 days (output): 5.52B (25.80M output)
Subscription $ / 1M output + input: $0.78
OpenRouter $ / 1M output + input: $1.23Morph
Subscription saving: 36.8% less

Tier / conditions: Off-peak rate; off-peak
Processed / 30 days (output): 27.62B (128.98M output)
Subscription $ / 1M output + input: $0.78
OpenRouter $ / 1M output + input: $1.23Morph
Subscription saving: 36.8% less

DeepSeek V4 Pro2 scenarios

Tier / conditions: Off-peak rate; off-peak
Processed / 30 days (output): 1.13B (5.27M output)
Subscription $ / 1M output + input: $3.80
OpenRouter $ / 1M output + input: $6.25Baidu, FP8
Subscription saving: 39.3% less

Tier / conditions: Off-peak rate; off-peak
Processed / 30 days (output): 5.64B (26.35M output)
Subscription $ / 1M output + input: $3.80
OpenRouter $ / 1M output + input: $6.25Baidu, FP8
Subscription saving: 39.3% less

Command Code8 recorded models / aliasesGo $1 / GOAT $10 / Pro $20 / Max 10x $100 / Max 20x $200 / month

Allowances are alternatives within one pool. Temporary deals (MiniMax M3 2x, Grok 4.7 40% off to Sep 27, DeepSeek V4.1 Flash boost to Sep 28, MiMo V2.5 up to 99% off) are excluded.

DeepSeek V4 Flash5 scenarios

Tier / conditions: Off-peak rates; peak doubles; off-peak
Processed / 30 days (output): 920.77M (4.30M output)
Subscription $ / 1M output + input: $0.23
OpenRouter $ / 1M output + input: $1.23Morph
Subscription saving: 81.0% less

Tier / conditions: Off-peak rates; peak doubles; off-peak
Processed / 30 days (output): 5.52B (25.80M output)
Subscription $ / 1M output + input: $0.39
OpenRouter $ / 1M output + input: $1.23Morph
Subscription saving: 68.4% less

Tier / conditions: Off-peak rates; peak doubles; off-peak
Processed / 30 days (output): 6.45B (30.10M output)
Subscription $ / 1M output + input: $0.66
OpenRouter $ / 1M output + input: $1.23Morph
Subscription saving: 45.8% less

Tier / conditions: Off-peak rates; peak doubles; off-peak
Processed / 30 days (output): 13.81B (64.49M output)
Subscription $ / 1M output + input: $1.55
OpenRouter $ / 1M output + input: $1.23Morph
Subscription saving: 26.4% more expensive

Tier / conditions: Off-peak rates; peak doubles; off-peak
Processed / 30 days (output): 27.62B (128.98M output)
Subscription $ / 1M output + input: $1.55
OpenRouter $ / 1M output + input: $1.23Morph
Subscription saving: 26.4% more expensive

GLM-5.2$9.75 subscription

Tier / conditions: GOAT
Processed / 30 days (output): 219.66M (1.03M output)
Subscription $ / 1M output + input: $9.75
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Kimi K2.7 Code$8.35 subscription

Tier / conditions: GOAT
Processed / 30 days (output): 256.39M (1.20M output)
Subscription $ / 1M output + input: $8.35
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

MiMo V2.6 Flash$0.95 subscription

Tier / conditions: GOAT
Processed / 30 days (output): 2.27B (10.58M output)
Subscription $ / 1M output + input: $0.95
OpenRouter $ / 1M output + input: $1.99Xiaomi, FP8
Subscription saving: 52.6% less

Kimi K3$49.45 subscription

Tier / conditions: GOAT
Processed / 30 days (output): 43.31M (0.20M output)
Subscription $ / 1M output + input: $49.45
OpenRouter $ / 1M output + input: $73.04Relace, FP4
Subscription saving: 32.3% less

GLM-5.3$34.12 subscription

Tier / conditions: GOAT
Processed / 30 days (output): 62.76M (0.29M output)
Subscription $ / 1M output + input: $34.12
OpenRouter $ / 1M output + input: $27.28Morph
Subscription saving: 25.1% more expensive

Claude Sonnet 5$66.64 subscription

Tier / conditions: Pro and above; $20 per premium model on Pro
Processed / 30 days (output): 64.28M (0.30M output)
Subscription $ / 1M output + input: $66.64
OpenRouter $ / 1M output + input: $70.31Anthropic
Subscription saving: 5.2% less

Claude Opus 5.52 scenarios

Tier / conditions: Max plans only; premium-model limit
Processed / 30 days (output): 232.47M (1.09M output)
Subscription $ / 1M output + input: $92.12
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Tier / conditions: Max plans only; premium-model limit
Processed / 30 days (output): 464.95M (2.17M output)
Subscription $ / 1M output + input: $92.12
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

StepFun Step Plan4 recorded models / aliasesFlash Mini $6.99 / Plus $9.99 / Pro $29 / Max $99 / month

The provider states the USD conversion as approximate. Credits reset monthly and do not roll over.

Step 3.5 Flash4 scenarios

Tier / conditions: Flash Mini
Processed / 30 days (output): 2.37B (11.09M output)
Subscription $ / 1M output + input: $0.63
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Tier / conditions: Flash Plus
Processed / 30 days (output): 9.50B (44.34M output)
Subscription $ / 1M output + input: $0.23
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Tier / conditions: Flash Pro
Processed / 30 days (output): 47.48B (221.72M output)
Subscription $ / 1M output + input: $0.13
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Tier / conditions: Flash Max
Processed / 30 days (output): 237.42B (1.11B output)
Subscription $ / 1M output + input: $0.09
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Step 5 Preview4 scenarios

Tier / conditions: Flash Mini
Processed / 30 days (output): 600.51M (2.80M output)
Subscription $ / 1M output + input: $2.49
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Tier / conditions: Flash Plus
Processed / 30 days (output): 2.40B (11.22M output)
Subscription $ / 1M output + input: $0.89
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Tier / conditions: Flash Pro
Processed / 30 days (output): 12.01B (56.08M output)
Subscription $ / 1M output + input: $0.52
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Tier / conditions: Flash Max
Processed / 30 days (output): 60.05B (280.39M output)
Subscription $ / 1M output + input: $0.35
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Step 3.7 FlashPricing not established

Tier / conditions: See plan terms
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

step-router-v1Pricing not established

Tier / conditions: Automatic routing
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

StepFun Step Plan (China)2 recorded models / aliasesFlash Mini ¥49 / Plus ¥99 / Pro ¥199 / Max ¥699 / month

China-region edition; CNY prices converted for comparison only.

Step 3.5 Flash4 scenarios

Tier / conditions: Flash Mini
Processed / 30 days (output): 2.37B (11.09M output)
Subscription $ / 1M output + input: $0.65
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Tier / conditions: Flash Plus
Processed / 30 days (output): 9.50B (44.34M output)
Subscription $ / 1M output + input: $0.33
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Tier / conditions: Flash Pro
Processed / 30 days (output): 47.48B (221.72M output)
Subscription $ / 1M output + input: $0.13
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Tier / conditions: Flash Max
Processed / 30 days (output): 237.42B (1.11B output)
Subscription $ / 1M output + input: $0.09
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Step 5 PreviewPricing not established

Tier / conditions: See plan terms
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

GLM Coding Plan (China)2 recorded models / aliasesLite ¥118 / Pro ¥538 / Max ¥1,078 / month

Sep 25 - Oct 7 all-day off-peak promotion excluded.

GLM-5.36 scenarios

Tier / conditions: Lite; off-peak
Processed / 30 days (output): 432.12M (2.02M output)
Subscription $ / 1M output + input: $8.66
OpenRouter $ / 1M output + input: $27.28Morph
Subscription saving: 68.2% less

Tier / conditions: Lite; peak
Processed / 30 days (output): 216.06M (1.01M output)
Subscription $ / 1M output + input: $17.33
OpenRouter $ / 1M output + input: $27.28Morph
Subscription saving: 36.5% less

Tier / conditions: Pro; off-peak
Processed / 30 days (output): 2.59B (12.11M output)
Subscription $ / 1M output + input: $6.59
OpenRouter $ / 1M output + input: $27.28Morph
Subscription saving: 75.9% less

Tier / conditions: Pro; peak
Processed / 30 days (output): 1.30B (6.05M output)
Subscription $ / 1M output + input: $13.17
OpenRouter $ / 1M output + input: $27.28Morph
Subscription saving: 51.7% less

Tier / conditions: Max; off-peak
Processed / 30 days (output): 6.05B (28.25M output)
Subscription $ / 1M output + input: $5.65
OpenRouter $ / 1M output + input: $27.28Morph
Subscription saving: 79.3% less

Tier / conditions: Max; peak
Processed / 30 days (output): 3.02B (14.12M output)
Subscription $ / 1M output + input: $11.31
OpenRouter $ / 1M output + input: $27.28Morph
Subscription saving: 58.5% less

GLM-5.3-FlashPricing not established

Tier / conditions: See plan terms
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

MiniMax Token Plan (China)2 recorded models / aliasesPlus ¥49 / Max ¥119 / Ultra ¥469 / month

China-region edition.

MiniMax M3$13.30 OpenRouter

Tier / conditions: All listed tiers
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: $13.30GMICloud, FP8
Subscription saving: Not established

MiniMax M2.7Pricing not established

Tier / conditions: See plan terms
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Alibaba Bailian Coding Plan (China)1 recorded model / aliasPro ¥200 (first month ¥39.90) / month

Lite closed to new purchase 2026-03-20; Pro sold in limited daily slots.

qwen3.7-plusPricing not established

Tier / conditions: Pro
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

Alibaba Bailian Token Plan (China)3 recorded models / aliasesLite ¥39 / Essential ¥79 / Standard ¥139 / Pro ¥499 (limited-time; list ¥60 / ¥120 / ¥180 / ¥600) / month

Team edition seats: ¥150 / ¥550 / ¥1,398 per seat (25k / 100k / 250k Credits).

qwen3.8-maxPricing not established

Tier / conditions: All personal tiers
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

deepseek-v4.1-flashPricing not established

Tier / conditions: Half Credits 22:00-08:00 (limited time)
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

glm-5.3Pricing not established

Tier / conditions: See plan terms
Processed / 30 days (output): Not established
Subscription $ / 1M output + input: Not established
OpenRouter $ / 1M output + input: Not quoted
Subscription saving: Not established

The selected route is the lowest modeled cost in the captured endpoints with status 0, advertised tool support and explicit cache-read pricing; when two eligible routes cost the same, the model author’s route is kept. Nonzero-status routes are excluded from selection. On September 25 every author route was eligible, and for MiMo and Sonnet it was also the cheapest. These are dated quotes, not a promise of availability or default routing. The snapshot records provider tags, context limits, quantization, rate overrides and source URLs in the price dataset.

Headline input prices can pick the wrong provider. For DeepSeek V4 Pro, Ionstream quotes $0.2528/M fresh input, below Baidu’s $0.3498/M. But its cache-read price is $0.08832/M versus Baidu’s $0.01113/M. Including their output rates and the purchase fee, this measured workload costs $23.21 through Ionstream versus $6.25 through Baidu. The quote, provider tag and context limit for that example are retained in the dataset. DeepSeek V4 Pro provider quotes

Reproducing the measured 96.53% input cache-hit rate is an assumption. OpenRouter uses sticky routing to help preserve caches, but fallback can move requests to a different provider. Verify actual billed cache reads in the coding harness. Pinning a provider or quantization changes the available routes; an FP4 quote is not evidence of equal application quality to an FP8 or author-hosted route. Each request must also fit that endpoint’s context and pricing tier. Caching, provider routing

Z.ai’s GLM-5.3-Flash author quote is back at its list price; the half-price promotion ended September 9. DeepSeek comparisons use OpenRouter’s V4.1 Flash and Pro 0813 entries; Chutes still serves the pinned V4 Flash 0731 checkpoint, and a subscription’s generic alias may not guarantee a specific checkpoint. The author-route DeepSeek quotes are off-peak. Several selected routes are FP4 or do not disclose quantization, including GLM-5.3, GLM-5.3-Flash, Kimi K3 and DeepSeek V4.1 Flash. Sonnet uses standard service and five-minute cache creation. Recheck these conditions before extending the estimates beyond this snapshot.

At full modeled utilization, OpenCode Go’s MiMo allocation buys about 6.33× the tokens per dollar of the selected OpenRouter route, and its MiniMax M3 allocation about 5.06×. Z.ai Max’s off-peak GLM-5.3 allocation buys about 4.59×, down from 7.73× on September 7 because a much cheaper GLM-5.3 route appeared. Two subscription allocations now lose to OpenRouter outright: OpenCode Go’s GLM-5.3 allowance costs 66.8% more than the selected route, and a Synthetic pack spent on GLM Flash costs about 7.5% more than InferenceNet’s FP4 route. Chutes Pro’s conditional DeepSeek Flash ceiling also remains more expensive than the selected route. These are different serving configurations, so accepted-work testing still decides whether the cheaper tokens help.

OpenRouter is therefore a useful baseline and overflow source. A subscription only beats it when enough relevant allowance is consumed: Go’s MiMo allocation breaks even at 15.8% utilization and Z.ai Max’s GLM-5.3 allocation at 21.8%, while Synthetic’s Kimi K3 example needs about 51.8% against the current selected quote. The OpenRouter comparison CSV includes cash-budget token capacity; the subscription comparison CSV includes savings and break-even utilization for every mapped, quantifiable offer.

Convert the allowance before comparing it#

Z.ai illustrates why a flat subscription cannot be reduced to one universal token price. Credits depend on the selected model, uncached input, cached input, output and time of day. Off-peak model usage consumes half the normal credits.

Applying its credit formula to the example workload gives these fully utilized, off-peak, steady-state 30-day equivalents:

Lite7.7× API value

Monthly fee: $18
GLM-5.3 output + input: 2.0M + 430M
Direct API equivalent: $138
Value multiple: 7.7×

Pro10.3× API value

Monthly fee: $80
GLM-5.3 output + input: 12.1M + 2581M
Direct API equivalent: $826
Value multiple: 10.3×

Max11.5× API value

Monthly fee: $168
GLM-5.3 output + input: 28.2M + 6021M
Direct API equivalent: $1,928
Value multiple: 11.5×

The weekly allowance is prorated for comparison; it is not a guaranteed calendar-month entitlement. Five-hour limits, concurrency, tool consumption and uneven demand can reduce realized value. All-peak operation halves the token allowances. The same subscriptions produce approximately 2.6–4.0× ordinary GLM Flash API value in this scenario, because Flash’s direct API is already much cheaper. Z.ai’s own documentation describes the monthly quota as roughly 15–30× the fee at API prices; this measured, cache-heavy workload converts to 7.7–11.5× for GLM-5.3. The provider does not state the workload behind its figure. Its subscription page renders prices in the browser, so the $18 / $80 / $168 fees come from the credits-based price set in the official site’s own code, consistent with the documentation’s “starting at just 18 USD per month”. Plan update notice

OpenCode Go has a different mechanism. Some models receive six times the subscription fee in API allowance; others receive three or one-and-a-half times. Allocating its entire qualifying allowance to MiMo V2.6 Flash supports approximately 31.7M output plus 6.76B input tokens for $10 in this scenario. Allocating it to MiniMax M3 supports approximately 3.8M output plus 811M input. Those are alternative uses of one pool, still subject to shorter limits. Each model is also capped at 20% of its monthly limit per five hours and 50% per week. Allowance rules

MiMo’s own plan shows the danger of reading credits as tokens. Each uncached V2.6 Flash input token costs 100 credits; each output token costs 200. V2.6 deducts exactly as V2.5 did, and V2.5 is deprecated on October 21. Its 4.1B-credit Lite plan therefore represents approximately $5.74 of daytime direct API usage for a $6 fee, or $7.18 with the off-peak credit discount. Large numbers need units. Conversion rules, API prices

Synthetic’s $24 weekly allowance represents about $103 per 30 days at Synthetic’s rates. Always compare the serving provider’s cache prices and model configuration with the original API before calling that a direct-provider discount. Limits, model metadata and prices

NanoGPT illustrates the effect of counting cache hits at full input weight. At the measured ratio, its 60M weekly input-unit allowance yields approximately 258M total processed tokens, including 1.21M output, per 30 days on a 1× model. A 2× model halves those estimates. A request-count plan needs another measurement—tokens per billable request—before it can be converted. The local message count is not assumed to equal provider API calls. NanoGPT quota definitions

Utilization is the other half of the arithmetic. A $10 plan containing $60 of relevant API usage breaks even once it replaces more than $10 of API spending. The unused $50 has no value. More packs can buy parallelism, but idle parallelism is still a bill.

Native coding subscriptions belong in the trial too#

Codex through ChatGPT, Claude Code, Cursor, Copilot, Gemini CLI, Devin and Grok Build can be economical when their native workflows produce useful results. Each now has its own row in the comparison above, with the provider’s own prices, models and limits. Only Copilot converts to token capacity from published numbers. Its credits are spent at listed per-token model prices. The others publish request counts, message estimates, relative multiples or undisclosed quotas. Those do not establish a cost per token, and the rows leave the capacity as not established rather than borrowing someone else’s measurement.

Temporary offers, kept separate#

As checked September 25, 2026:

Z.ai charges all-day usage at the off-peak rate from September 25 to October 7. Its GLM-5.3-Flash campaign, extended to October 7, gives unlimited Flash use through ZCode and AutoClaw daily 15:00–01:00 UTC and doubled quota through other supported agents in that window. Plan notice, Flash campaign
OpenCode Go gives DeepSeek V4.1 Flash four times its $15 allowance, $60, until September 27, and lists Space Bunny as free for a limited time. Go usage
deepseekv4pro.com advertises a 2× quota on its Coding tier with no stated end date. Pricing

These offers can improve a trial; the standing calculations exclude all of them. Z.ai’s half-price GLM-5.3-Flash API offer ended September 9 and has been removed from the price data.

The next analysis should measure accepted work#

Run 20–50 representative tasks through the same harness, tests and retry budget. Record input, cache hits, billed output, subscription consumption, cash spend, retries, queue time, review time and whether the change was accepted. Include abandoned attempts in the bill.

The metric is total inference and retry cost divided by changes that pass tests and review. Report elapsed time alongside it: a cheap queue can become expensive when it delays everything else. Allocate the full subscription fee across the measured period, including unused quota, rather than claiming the theoretical maximum discount.

That experiment may favor a cheap default with occasional stronger calls. It may favor one reliable model that finishes sooner. This initial research does not establish which outcome the fleet will produce.

Reproduce and update the comparison#

Download the dated price dataset, measured usage profile, model catalogs, calculator, API comparison CSV, plan capacity CSV, Z.ai comparison CSV, OpenRouter quote CSV, and OpenRouter versus subscriptions CSV. The dataset records source URLs and check dates; the artifact’s comparisons and CSVs are generated from it.

Put calculate.mjs and prices.json in the same directory. With Node 22 or newer:

node calculate.mjs --format api > api.csv
node calculate.mjs --format plans > capacity.csv
node calculate.mjs --format zai > zai.csv
node calculate.mjs --format openrouter > openrouter.csv
node calculate.mjs --format openrouter-plans > openrouter-plans.csv
# Sensitivity scenario; defaults above use the measured profile.
node calculate.mjs --ratio 50 --cache-hit 0.95 --cache-write-share 0.2 --format plans

--cache-write-share is the fraction of non-cache-read input spent creating caches. It preserves separate write pricing while --cache-hit controls reads as a fraction of all input. The calculator does not convert a message count into requests.

For OpenRouter, --topup-credits 10 models the fee on a $10 inference-credit purchase; --budget remains a cash budget. The default amortizes a $100 credit purchase. Ratio overrides reprice the captured provider routes; they do not select a new cheapest provider. Use the recorded endpoint API URLs to refresh route selection for a different workload.

Future revisions can add measured task outcomes, new providers, different cache-hit scenarios and throughput observations without moving this artifact. The page-level verification date advances only after the comparison has been checked as a whole; partial refreshes retain individual source dates. Corrections and changed recommendations belong in the changelog, newest first.

Changelog#

2026-09-25 — Every tracked subscription, first-party only#

Added fifteen subscriptions, bringing the comparison to twenty-seven: ChatGPT (Codex), Claude (Claude Code), Cursor, GitHub Copilot, Google AI Pro / Ultra, Devin, SuperGrok, Ollama Cloud, Command Code, StepFun Step Plan (international and China), and China-region editions of the GLM Coding Plan, MiniMax Token Plan, Alibaba Bailian Coding Plan and Alibaba Bailian Token Plan.
Calculated capacity only where the provider publishes both an allowance and the rates it is spent at: Copilot, Ollama Cloud, Command Code, StepFun and the GLM China edition. Every other new row quotes the provider’s own limits and leaves capacity unestablished; no third-party usage measurement was adopted.
Converted CNY prices at the CFETS central parity of 6.7489 CNY/USD for September 24, recorded in the dataset. Kimi’s China-region price list is served only inside China, so it is noted on the existing Kimi row rather than added.
Recorded two first-party discrepancies with other trackers. Devin’s own pricing page ends the SWE-2 unmetered promotion on October 10. SuperGrok Lite and Heavy prices are not published on xAI’s pricing page, so they are not shown.
The native-subscription section now points to these rows instead of summarizing them.

2026-09-25 — First-party re-verification#

Rechecked every plan, direct API price, OpenRouter route and native subscription against the providers’ own pages and APIs, and advanced the page-level verification date. The September 7 usage measurement is unchanged.
Every adopted value is now first-party evidence: the provider’s own published page or API, or my own measurement. An independent third-party dataset was used only to find what to recheck; none of its values are adopted. Where a provider publishes no convertible allowance (MiniMax, Kimi, Alibaba, deepseekv4pro.com), the plan stays unquantified and its note records what the provider does publish.
MiMo: priced V2.6 Flash and V2.6 Pro, which bill and deduct exactly as V2.5 did; V2.5 is deprecated October 21.
DeepSeek: the V4 Flash names are retired and billed at V4.1 Flash off-peak rates of $0.15 / $0.003 / $0.60, down from $0.22 / $0.007 / $0.66. OpenCode Go’s Flash allocation was recalculated at those rates.
OpenCode Go: added the GPT 6 Luna allocation and the current per-model limits and catalog (Grok 4.7, MiMo V2.6, DeepSeek V4.1 Flash, MiniMax M2.5, Space Bunny; Omen Alpha removed).
Synthetic withdrew GLM-5.2 and added DeepSeek V4.1 Flash. Updated the camelStream fleet, Kimi tiers and quota rules, Alibaba’s limits, Chutes’ overage discounts, the NanoGPT and Kilo catalogs, and deepseekv4pro.com’s prices and plan-token quotas.
Re-captured OpenRouter with the same selection rule, adding an author-route tie-break. Z.ai Max’s GLM-5.3 advantage fell from 7.73× to 4.59×, and OpenCode Go’s GLM-5.3 and Synthetic’s GLM Flash allocations now cost more than the cheapest eligible route. The MiMo cache-price example no longer held and was replaced with a DeepSeek V4 Pro example.
Removed the expired GLM-5.3-Flash promotional price and rewrote the temporary-offers section.

2026-09-07 — Published as a data artifact#

Retired the private Notes draft and published the maintained comparison at /data/inference-plans/.
Kept the report, Markdown twin, calculator, source data and generated CSVs together under one public artifact path.
This changes the publication surface only; the analysis, measurement and source-verification dates are unchanged.

2026-09-07 — Display the analysis date#

Added a prominent analysis date near the top of the artifact, separate from publication, measurement and source-verification dates.

2026-09-07 — Index by subscription and model#

Reorganized the OpenRouter comparison into all twelve researched coding subscriptions, then their models, with tier and time-window rows underneath.
Included the saved model catalogs; large unpriced catalogs expand within the relevant subscription instead of disappearing from the comparison.
Added token capacity beside each calculated model/tier cost and retained explicit unknowns for missing allowances, rates and model multipliers.
Preserved the existing measured mix and dated prices; this is a coverage and presentation update, not a source refresh.

2026-09-07 — OpenRouter route comparison#

Added nine model comparisons using coherent provider quotes, measured cache composition and the standard credit-purchase fee.
Added selected provider, quantization, context and status metadata; separated flagged author quotes and current promotional rates.
Added OpenRouter cash-budget capacity and subscription savings/break-even CSVs, plus a configurable top-up size.
Found that provider cache pricing can reverse headline-price rankings, and that the selected DeepSeek Flash route undercuts Chutes Pro’s modeled ceiling.
Retained the earlier usage measurement and subscription snapshot; this revision refreshes OpenRouter quotes only.

2026-09-07 — Normalize subscription costs#

Converted the measured mix into effective costs per million processed tokens and per million output tokens including associated input for every quantifiable offer.
Added a main normalization table, expandable model/time-window calculations and matching CSV fields.
Added explicit same-model direct API savings and break-even utilization, preserving negative savings and unknown comparisons.
Kept the existing measurement and dated price snapshot; this revision adds arithmetic rather than claiming a new measurement or source refresh.

2026-09-07 — Measured usage, supported models and capacity#

Replaced the illustrative 20:1 input/output ratio and 80% cache-hit assumption with a fresh local measurement: 213.2:1 and 96.53% of input from cache.
Normalized separately reported reasoning into billable output and retained distinct cache-read and cache-write buckets.
Added model coverage for every plan in the main table, dated catalog snapshots and reproducible token-capacity estimates using serving-provider rates.
Kept unknown quotas, unmetered service and conditional ceilings explicit; added downloadable usage and capacity data.
Recalculated API examples. The shortlist remains a set of trial candidates; no accepted-work benchmark has yet been run.

2026-09-07 — Initial research draft#

Compared twelve subscription offerings with direct API costs and native coding alternatives.
Added reproducible API and Z.ai calculations, source-check dates, and explicit workload assumptions.
Recorded automation restrictions, conflicting reseller documentation and expiring promotions.
Established the cost-per-accepted-change experiment as the next analysis; no paid performance benchmark has been run for this note.