Data API·MCP
Build on the most complete
LLM intelligence layer
Query benchmark scores, model rankings, pricing, and metadata across every model we track — over the REST API, or directly from your AI agent over MCP.
/rankingscoding
1
claude-4-opusAnthropic
94.2
2
gpt-5OpenAI
93.1
3
gemini-2.5-proGoogle
91.8
4
llama-4-maverickMeta
89.4
5
deepseek-r2DeepSeek
87.6
/models
claude-4-sonnetAnthropic
200Ktextvision
$3/M in·$15/M out
gpt-5-miniOpenAI
128Ktext
$0.40/M in·$1.60/M out
gemini-2.5-flashGoogle
1Mtextvisionaudio
$0.15/M in·$0.60/M out
/scores
HumanEval
MMLU-Pro
MATH-500
GPQA
claude-4-opus
96.3
89.2
97.1
72.4
gpt-5
94.8
91.5
95.3
68.7
gemini-2.5-pro
93.1
88.7
93.8
70.2
MCP server
MCP docs/mcp
hostedConnect your AI agent to the LLM Stats catalog over MCP. The same ze_… key authenticates the REST API, the inference gateway, and the MCP server.
https://mcp.llm-stats.com/mcpTools11
list-models
Filter the catalog by org, modality, price, context
get-model
Full detail for one model — scores, pricing, sources
list-benchmarks
Benchmark catalog with categories and model counts
list-scores
Score matrix across models and benchmarks
get-rankings
TrueSkill rankings for any category
list-updates
Models added in the last N days
compare-models
Side-by-side comparison of 2–4 model ids
find-best-models-for-category
Top-ranked models, optionally filtered by price
resolve-model
Fuzzy lookup: free text → ranked candidate ids
compare-run
Run one prompt through 2–4 models in parallel
get-gateway-snippet
Paste-ready curl, Python, or JS snippet
read
workflow
gateway
Slash-commands
/compare-models/find-cheapest-for/whats-new-this-week/explain-benchmark
InstallAdd to .cursor/mcp.json
{
"mcpServers": {
"llm-stats": {
"url": "https://mcp.llm-stats.com/mcp",
"headers": {
"Authorization": "Bearer ze_<your-api-key>"
}
}
}
}Replace ze_<your-api-key> with the key from your developer console.
REST endpoints
GET
/stats/v1/modelsCatalog with metadata, pricing, and category scoresGET
/stats/v1/models/{id}Full model detail with every benchmark scoreGET
/stats/v1/benchmarksAll benchmarks with categories and model countsGET
/stats/v1/scoresScore matrix — filter across models and benchmarksGET
/stats/v1/rankingsTrueSkill rankings by categoryGET
/stats/v1/updatesRecently added models (1–30 day lookback)Ready to build?
One ze_… key unlocks the REST API and the MCP server. Sign in to mint yours in a minute.