Benchmarks

Compare models within one benchmark. Scores are source-linked, not an overall ranking.

154 Benchmarks
1,324 Model routes
13,002 Route rows

Sources include model cards, papers, and benchmark operators. Catalog updated Nov 16, 2026.

Keeps the selected benchmark and clears search, source, confidence, and sort filters.
SWE-bench VerifiedMMLUGPQA DiamondAider PolyglotMistral 7B comparison tableHumanEvalArtificial Analysis Coding IndexArtificial Analysis Intelligence Index

Benchmark results

Results are grouped by the selected benchmark.

Loading...

Family Score Metric Category Scope Routes Source
Loading benchmark rows...