What is the LMArena Agent benchmark?
LMArena's Agent board evaluates models running multi-step agentic sessions with tools; scores are the published win share with its official interval, per evaluated configuration. It is published by LMArena.
Benchmark leaderboard
LMArena's Agent board evaluates models running multi-step agentic sessions with tools; scores are the published win share with its official interval, per evaluated configuration. This page lists every model ModelCap tracks with a published LMArena Agent result, its best evaluated configuration, the source's own rank, and the model's ModelCap Index position, API price and context window.
Snapshot as of 18 November 2026
Claude Fable 5.2 from Anthropic leads the LMArena Agent leaderboard among the 56 tracked models with 27.4% (Max configuration), per the source snapshot published 18 November 2026.
Sorted by published score, one row per model. Source: LMArena.
LMArena Agent ranks models by agentic session score. ModelCap ingests the board as published, matches each entry to a catalogue model with a reviewed identity, and shows the source's score, interval and rank unchanged. Not on its own. The ModelCap Index combines several public capability sources with published uncertainty; LMArena Agent contributes as agent evidence where a model has an admitted result. The methodology page documents the weighting. Read the methodology and identity rules.
LMArena's Agent board evaluates models running multi-step agentic sessions with tools; scores are the published win share with its official interval, per evaluated configuration. It is published by LMArena.
Claude Fable 5.2 (Anthropic) holds the top LMArena Agent score among the models ModelCap tracks, at 27.4% as of the source snapshot published 18 November 2026.
56 tracked models have a published LMArena Agent result on ModelCap; the source board itself lists 68 entries. Each row shows the model's best evaluated configuration.
Inkling 2 Small from Thinking Machines is the highest-scoring model with openly downloadable weights on this board, at 1.5%.
Among the ten highest-scoring priced models, Gemini 4 Hydrogen has the lowest listed output price at $7.50 per 1M tokens while scoring 13.5%.
The board ranks models by agentic session score; higher scores are better. ModelCap shows the source's own score, interval and rank and never re-runs the evaluation.
Not on its own. The ModelCap Index combines several public capability sources with published uncertainty; LMArena Agent contributes as agent evidence where a model has an admitted result. The methodology page documents the weighting.
The newest LMArena Agent publication ModelCap holds is dated 18 November 2026. This page uses the published snapshot; source collection, publication and page updates happen separately.