The LLM data API

Benchmark scores, rankings, pricing, and metadata for every model — over REST or MCP.

Free·Unlimited requests

Used by people at

OpenAI
Google
Meta
Amazon
IBM
Reddit
Stanford
Harvard
UC Berkeley
Tsinghua
University of Hong Kong
Hugging Face
Y Combinator
Xataka
TechCrunch
Seeking Alpha
Datadog
NBC News
The Lancet
OpenAI
Google
Meta
Amazon
IBM
Reddit
Stanford
Harvard
UC Berkeley
Tsinghua
University of Hong Kong
Hugging Face
Y Combinator
Xataka
TechCrunch
Seeking Alpha
Datadog
NBC News
The Lancet
380+
Models tracked

Every major model with benchmark scores, pricing, context length, and provider metadata. Updated within hours of release.

50+
Verified benchmarks

Coding, math, reasoning, vision, and more. Scores are source-verified, not self-reported.

11
MCP tools

Connect any AI agent — Cursor, Claude Code, Windsurf — with one-click OAuth. No key to copy.

Request
curl -H "Authorization: Bearer $API_KEY" \
  "https://api.zeroeval.com/stats/v1/models?organization=anthropic&limit=2"
Response
{
  "models": [
    {
      "id": "claude-4-sonnet",
      "name": "Claude 4 Sonnet",
      "organization": { "id": "anthropic", "name": "Anthropic" },
      "providers": [{ "provider_name": "Anthropic", "input_price_per_m": 3.0 }],
      "top_scores": { "coding": 94.2 }
    }
  ],
  "next_cursor": "...",
  "total": 382
}
/stats/v1/modelsCatalog with metadata, pricing, and category scores
/stats/v1/models/{id}Full model detail with every benchmark score
/stats/v1/benchmarksAll benchmarks with categories and model counts
/stats/v1/scoresScore matrix — filter across models and benchmarks
/stats/v1/rankingsTrueSkill rankings by category
/stats/v1/updatesRecently added models (1–30 day lookback)

Start building

Completely free with no usage limits. Sign in to get your API key.