LLM Pricing API

Free, schema-validated JSON dataset of current per-token pricing, context-window size, and quality tier for the major 2026 LLMs. No API key, no rate limit, CC BY 4.0. Drop it into your own tool, dashboard, or AI agent.

Endpoints

GET https://inference-labs.com/api/prices.json
GET https://inference-labs.com/api/prices/{model-id}.json
GET https://inference-labs.com/api/prices.schema.json

OpenAPI spec

The full surface (free pricing + authenticated router) is described in /api/openapi.yaml — OpenAPI 3.1, importable into Postman, Insomnia, Bruno, Swagger UI, or any code generator.

Quick example

curl -sS https://inference-labs.com/api/prices.json \
  | jq '.models[] | select(.tier=="cheap") | {id, input: .price_per_million.input_usd, output: .price_per_million.output_usd}'

Per-model endpoints

/api/prices/gpt-5.json — GPT-5
/api/prices/gpt-5-mini.json — GPT-5 mini
/api/prices/gpt-4o-mini.json — GPT-4o mini
/api/prices/claude-sonnet-4-5.json — Claude Sonnet 4.5
/api/prices/claude-haiku-4-5.json — Claude Haiku 4.5
/api/prices/claude-opus-4-7.json — Claude Opus 4.7
/api/prices/gemini-2-5-pro.json — Gemini 2.5 Pro
/api/prices/gemini-2-5-flash.json — Gemini 2.5 Flash
/api/prices/nova-pro.json — Amazon Nova Pro
/api/prices/llama-3-3-70b.json — Llama 3.3 70B

Embeddable price table

Drop this into any README or docs page:

<iframe src="https://inference-labs.com/embed/price-table.html"
        width="100%" height="520" frameborder="0"></iframe>

Live preview: /embed/price-table.html

Badges

shields.io-style SVG badges that always show current values:

https://inference-labs.com/badge/cheapest-llm.svg
https://inference-labs.com/badge/largest-context.svg
https://inference-labs.com/badge/cheapest-1m-context.svg
https://inference-labs.com/badge/best-value.svg

License

CC BY 4.0. Use the data in any project (commercial or otherwise). A link back to inference-labs.com is appreciated.

Updates

The dataset is regenerated whenever provider pricing changes (currently manual; check the updated field). For a hosted router that always uses the latest pricing internally, see the Inference Labs platform.