FINE PRINT FOR LLM APIs · CHECKED 2026-10-22
LLM API prices, with the fine print
What a call will actually cost, per host, region and mode, with the vendor's own words and the date on every number. Plus retirement dates, replacements, data terms, and where the popular free price lists disagree with the vendor (18 of 274 checks today).
Standard mode, first-party API, per 1M tokens. Bedrock, Vertex, batch, flex, priority, long-context and cache-write prices are in the API. Unknown is shown as —, never as $0.
OpenAI
| Model | Status | Context | Input | Cached in | Output | Retires |
|---|
| gpt-6-astra | ga | 1.05M | $10 | $1 | $50 | — |
| gpt-6.1-sol | ga | 1.05M | $2 | $0.1 | $10 | — |
| gpt-6-sol | ga | 1.05M | $2 | $0.2 | $10 | — |
| gpt-6-luna | ga | 1.05M | $0.1 | $0.01 | $0.5 | — |
| gpt-5.6-sol | ga | 1.05M | $4 | $0.4 | $20 | — |
| gpt-5.6-terra | ga | 1.05M | $2 | $0.2 | $12 | — |
| gpt-5.6-luna | ga | 1.05M | $0.2 | $0.02 | $1.2 | — |
| gpt-5.1 | deprecated | 400K | $1.25 | $0.125 | $10 | 2027-04-01 |
| gpt-5.4-nano | deprecated | 400K | $0.2 | $0.02 | $1.25 | 2027-04-01 |
| gpt-4.1 | deprecated | 1.05M | $2 | $0.5 | $8 | 2027-01-14 |
| gpt-4o | deprecated | 128K | $2.5 | $1.25 | $10 | 2027-01-14 |
| o4-mini | deprecated | 200K | $1.1 | $0.275 | $4.4 | 2026-10-23 |
anthropic
| Model | Status | Context | Input | Cached in | Output | Retires |
|---|
| claude-fable-5-1 | ga | 1M | $10 | $0.25 | $50 | — |
| claude-opus-5-5 | ga | 1M | $4 | $0.2 | $20 | — |
| claude-sonnet-5-5 | ga | 1M | $2 | $0.1 | $10 | — |
| claude-haiku-5-5 | ga | 1M | $0.1 | $0.01 | $0.5 | — |
| claude-sonnet-4-5-20250929 | deprecated | 200K | $3 | $0.3 | $15 | 2026-11-30 |
| claude-haiku-4-5-20251001 | ga | 200K | $1 | $0.1 | $5 | — |
| claude-opus-4-5-20251101 | ga | 200K | $5 | $0.5 | $25 | — |
google
| Model | Status | Context | Input | Cached in | Output | Retires |
|---|
| gemini-4-argon | ga | 1.05M | $2 | $0.1 | $10 | — |
| gemini-3.1-pro-preview | preview | 1.05M | $2 | $0.2 | $12 | — |
| gemini-3.8-flash | ga | 1.05M | $0.75 | $0.075 | $3.75 | — |
| gemini-3.6-flash | ga | 1.05M | $0.75 | $0.075 | $3.75 | — |
| gemini-3.5-flash-lite | ga | 1.05M | $0.3 | $0.03 | $2.5 | — |
| gemini-3.1-flash-lite | ga | 1.05M | $0.25 | $0.025 | $1.5 | 2027-05-07 |
| gemini-3.7-flash | deprecated | 1.05M | — | — | — | — |
| gemini-3.5-flash | deprecated | 1.05M | — | — | — | — |
xAI
| Model | Status | Context | Input | Cached in | Output | Retires |
|---|
| grok-4.7 | unstated | 500K | $2 | $0.5 | $6 | — |
| grok-4.3 | unstated | 1M | $1.25 | $0.2 | $2.5 | — |
| grok-build-0.1 | unstated | 256K | $1 | $0.2 | $2 | — |
| grok-4-1-fast-reasoning | retired | — | — | — | — | 2026-05-15 |
mistral
| Model | Status | Context | Input | Cached in | Output | Retires |
|---|
| mistral-large-4 | preview | 1M | $1.36 | $0.14 | $4.18 | — |
| mistral-large-2512 | ga | 256K | $0.5 | — | $1.5 | — |
| mistral-medium-3-5 | ga | 256K | $1.5 | — | $7.5 | — |
| mistral-small-2603 | ga | 256K | $0.15 | — | $0.6 | — |
| zai-glm-5-2 | deprecated | 1M | $1.4 | $0.14 | $4.4 | 2026-10-31 |
DeepSeek
| Model | Status | Context | Input | Cached in | Output | Retires |
|---|
| deepseek-v4-pro | ga | 1M | $1.32 | $0.044 | $3.96 | — |
| deepseek-flash | unstated | 1M | $0.3 | $0.006 | $1.2 | — |
| deepseek-chat | retired | — | — | — | — | 2026-07-24 |
| deepseek-reasoner | retired | — | — | — | — | 2026-07-24 |
Coming up
- 2026-10-23 · o4-mini retires; replacement gpt-5.6-terra
- 2026-10-31 · zai-glm-5-2 retires; replacement zai-glm-5-3
- 2026-11-30 · claude-sonnet-4-5-20250929 retires; replacement claude-sonnet-5-5
- 2026-12-01 · gemini-4-argon on google-ai-studio: promotional price ends; new list price (input $4.0, output $20.0 per 1M tokens) across 12 rows
- 2027-01-01 · gemini-3.6-flash on google-ai-studio: promotional price ends; new list price (input $1.5, output $7.5 per 1M tokens) across 12 rows
- 2027-01-01 · gemini-3.8-flash on google-ai-studio: promotional price ends; new list price (input $1.5, output $7.5 per 1M tokens) across 12 rows
- 2027-01-14 · gpt-4o retires; replacement gpt-6-luna
- 2027-01-14 · gpt-4.1 retires; replacement gpt-6-luna
- 2027-04-01 · gpt-5.4-nano retires; replacement gpt-6-luna
- 2027-04-01 · gpt-5.1 retires; replacement gpt-6-sol
- 2027-05-07 · gemini-3.1-flash-lite retires; replacement gemini-3.5-flash-lite
For agents and builders
GET https://money-lab.app/api/llm/model?id=claude-sonnet-5-5
GET https://money-lab.app/api/llm/estimate?model=gemini-3.1-pro-preview&input_tokens=250000&output_tokens=2000
GET https://money-lab.app/api/llm/pick?min_context=500000&max_input_usd=1
GET https://money-lab.app/api/llm/check?id=o4-mini&fail_within_days=90 (HTTP 409 when it fails, for CI)
MCP: llm_model, llm_estimate_cost, llm_pick_model, llm_check_model, llm_changes on https://money-lab.app/api/mcp
Free: 50 calls a day per network with nothing attached, 500 with a free key. Price history ($0.005), full disagreement evidence ($0.01) and the whole dataset ($1) over x402.