Docs

Public pricing API

A free, read-only JSON API over the same 162-model, 11-provider dataset that powers the rest of Meterary. No authentication, no key, CORS-open — built to be fetched by scripts, dashboards and AI agents directly.

Endpoints

GET/api/v1/modelsEvery priced model. Optional ?provider=<slug> filter.

GET/api/v1/models/[slug]One model plus its documented price changes. 404 JSON on an unknown slug.

GET/api/v1/providersEvery provider, with its tracked model slugs nested under models.

Example

The response below is generated from the live dataset at build time, using the first model in it — it will always match whatever /api/v1/models actually returns.

curl https://meterary.com/api/v1/models/qwen-flash

{
  "data": {
    "slug": "qwen-flash",
    "name": "Qwen-Flash",
    "provider": "alibaba",
    "family": "Qwen",
    "apiId": "qwen-flash",
    "releaseDate": null,
    "status": "current",
    "modality": "text",
    "contextWindow": null,
    "maxOutput": null,
    "inputPerMTok": 0.05,
    "outputPerMTok": 0.4,
    "cachedInputPerMTok": null,
    "cacheWritePerMTok": null,
    "batchDiscountPct": null,
    "capabilities": [],
    "featured": false,
    "notes": "International pricing, base tier (0-256K); 256K-1M tier is $0.25/$2 per 1M tokens. Alibaba recommends this as the replacement for Qwen-Turbo.",
    "sourceUrl": "https://www.alibabacloud.com/help/en/model-studio/model-pricing",
    "verifiedAt": "2026-07-02",
    "confidence": "verified",
    "priceChanges": []
  }
}

The list endpoint wraps every model the same way, plus a meta block:

curl "https://meterary.com/api/v1/models?provider=anthropic"

{
  "data": [
    {
      "slug": "qwen-flash",
      "name": "Qwen-Flash",
      "provider": "alibaba",
      "family": "Qwen",
      "apiId": "qwen-flash",
      "releaseDate": null,
      "status": "current",
      "modality": "text",
      "contextWindow": null,
      "maxOutput": null,
      "inputPerMTok": 0.05,
      "outputPerMTok": 0.4,
      "cachedInputPerMTok": null,
      "cacheWritePerMTok": null,
      "batchDiscountPct": null,
      "capabilities": [],
      "featured": false,
      "notes": "International pricing, base tier (0-256K); 256K-1M tier is $0.25/$2 per 1M tokens. Alibaba recommends this as the replacement for Qwen-Turbo.",
      "sourceUrl": "https://www.alibabacloud.com/help/en/model-studio/model-pricing",
      "verifiedAt": "2026-07-02",
      "confidence": "verified"
    },
    "... more models"
  ],
  "meta": {
    "count": 162,
    "latestVerifiedAt": "2026-07-02",
    "source": "https://meterary.com",
    "license": "Free to use with attribution to Meterary (https://meterary.com)."
  }
}

Field reference

Every model object matches the PricedModel shape below (nullable fields mean the provider doesn't publish that number, or we couldn't verify it live).

Field	Type	Notes
slug	string	Kebab-case, unique across the dataset.
name	string	Official display name.
provider	string	Provider slug — see /api/v1/providers.
family	string	Model family label, e.g. "Claude", "GPT".
apiId	string \| null	The id passed to the provider's own API, if documented.
releaseDate	string \| null	ISO date, or null if undocumented.
status	"current" \| "legacy" \| "deprecated"	Provider-stated lifecycle stage.
modality	"text" \| "multimodal" \| "embedding" \| "image" \| "audio"	Primary I/O modality.
contextWindow	number \| null	Max input tokens.
maxOutput	number \| null	Max output tokens.
inputPerMTok	number \| null	USD per 1M input tokens, standard tier.
outputPerMTok	number \| null	USD per 1M output tokens, standard tier.
cachedInputPerMTok	number \| null	USD per 1M cached/read input tokens.
cacheWritePerMTok	number \| null	USD per 1M cache-write tokens.
batchDiscountPct	number \| null	Percent discount for async batch processing.
capabilities	string[]	e.g. "vision", "tool-use", "extended-thinking".
featured	boolean	Flagship model used on comparison pages.
notes	string?	Short factual note (tiered pricing, etc.).
sourceUrl	string	Official provider page the numbers were read from.
verifiedAt	string	ISO date the source page was actually fetched.
confidence	"verified" \| "verify"	"verify" means the source was unreachable at last check.
priceChanges	PriceChange[]	Only present on /api/v1/models/[slug].

Rate limits

No key, no hard cap — this is a static dataset served from cache. We ask for fair use: cache responses on your end for at least an hour (the Cache-Control header already tells any CDN or HTTP client to do this) rather than polling on every request.

Attribution

This dataset is free to use, including commercially, with attribution to Meterary (https://meterary.com) and a link back where practical. Every model also carries its own official sourceUrl — cite that too when you're quoting a specific price.

For AI agents

Start at https://meterary.com/llms.txt for a plain-text summary of this site and its key URLs, then call these endpoints directly for structured data — no HTML parsing required. The same facts are also embedded as a schema.org Dataset JSON-LD block on the /models directory page, so a crawler that only reads structured data still gets the full pricing table.