LMArena Agent Arena
LMArena · agent-2026-10-16-a64ca3b12b6e · Outcome-based measurements from real LMArena agent sessions.
Model Ranking
Source | |||||||
|---|---|---|---|---|---|---|---|
| #1 | Claude Fable 5.1Anthropic | 14.03 | — | 16,374 | Claude Fable 5.1 (Max); 95% CI [12.03932907, 16.02502568]; sessions 16374; observations 1913479; rank 3unknown | Original sourcewinner eligible | Third-party benchmark |
| #2 | Claude Opus 5Anthropic | 8.68 | — | 23,770 | Claude Opus 5 (Max); 95% CI [6.99592257, 10.36856205]; sessions 23770; observations 3146528; rank 8unknown | Original sourcewinner eligible | Third-party benchmark |
| #3 | Claude Fable 5Anthropic | 8.18 | — | 42,608 | Claude Fable 5 (High); 95% CI [6.75735259, 9.60649884]; sessions 42608; observations 2530600; rank 9unknown | Original sourcewinner eligible | Third-party benchmark |
| #4 | Claude Sonnet 5Anthropic | 4.58 | — | 33,276 | Claude Sonnet 5 (High); 95% CI [2.88280252, 6.28024714]; sessions 33276; observations 3570560; rank 12unknown | Original sourcewinner eligible | Third-party benchmark |
| #5 | Muse Spark 1.3Meta | 3.82 | — | 73,912 | Muse Spark 1.3 (Max); 95% CI [2.98809934, 4.65329623]; sessions 73912; observations 7381209; rank 14unknown | Original sourcewinner eligible | Third-party benchmark |
| #6 | Gemini 3.8 FlashGoogle DeepMindSelected model | 3.75 | — | 37,692 | Gemini 3.8 Flash (High); 95% CI [2.64957678, 4.84993702]; sessions 37692; observations 5316787; rank 15unknown | Original sourcewinner eligible | Third-party benchmark |
| #7 | Grok 4.7xAI | 3.21 | — | 15,127 | Grok 4.7 (xHigh); 95% CI [1.57579589, 4.84117305]; sessions 15127; observations 2475030; rank 16unknown | Original sourcewinner eligible | Third-party benchmark |
| #8 | Hy4 previewTencent | 2.68 | — | 84,828 | Hy4 preview; 95% CI [2.00929922, 3.35139547]; sessions 84828; observations 8314177; rank 20unknown | Original sourcewinner eligible | Third-party benchmark |
| #9 | Kimi K3Moonshot AI | 1.93 | — | 135,415 | Kimi K3 (Max); 95% CI [1.20183745, 2.66737332]; sessions 135415; observations 12343129; rank 23unknown | Original sourcewinner eligible | Third-party benchmark |
| #10 | Grok 4.6xAI | 0.89 | — | 28,868 | Grok 4.6 (xHigh); 95% CI [-0.22077562, 2.00905409]; sessions 28868; observations 3315294; rank 28unknown | Original sourcewinner eligible | Third-party benchmark |
| #11 | Gemini 3.7 FlashGoogle DeepMind | -0.72 | — | 54,989 | Gemini 3.7 Flash (High); 95% CI [-1.64184585, 0.19418642]; sessions 54989; observations 4923541; rank 32unknown | Original sourcewinner eligible | Third-party benchmark |
| #12 | Gemini 3.1 ProGoogle DeepMind | -4.00 | — | 87,236 | Gemini 3.1 Pro Preview; 95% CI [-5.35690873, -2.63856355]; sessions 87236; observations 2815597; rank 39unknown | Original sourcewinner eligible | Third-party benchmark |
| #13 | Hy3Tencent | -6.00 | — | 31,296 | Hy3; 95% CI [-7.50774599, -4.49280945]; sessions 31296; observations 1149668; rank 41unknown | Original sourcewinner eligible | Third-party benchmark |
| #14 | Muse Spark 1.2Meta | -6.41 | — | 51,435 | Muse Spark 1.2 (xHigh); 95% CI [-7.24140950, -5.58489868]; sessions 51435; observations 3268541; rank 43unknown | Original sourcewinner eligible | Third-party benchmark |
| #15 | Muse Spark 1.1Meta | -6.43 | — | 145,940 | Muse Spark 1.1; 95% CI [-7.08799618, -5.77544358]; sessions 145940; observations 7748331; rank 44unknown | Original sourcewinner eligible | Third-party benchmark |
| #16 | Mistral Medium 3.5 128BMistral AI | -11.72 | — | 12,762 | Mistral Medium 3.5; 95% CI [-13.50841359, -9.93416425]; sessions 12762; observations 527876; rank 49unknown | Original sourcewinner eligible | Third-party benchmark |
| #17 | InklingThinking Machines Lab | -14.35 | — | 43,507 | Inkling; 95% CI [-15.81859883, -12.88168979]; sessions 43507; observations 1662352; rank 50unknown | Original sourcewinner eligible | Third-party benchmark |
| #18 | Solar Pro 4Upstage | -20.06 | — | 10,243 | Solar Pro 4; 95% CI [-22.42702213, -17.68944380]; sessions 10243; observations 452978; rank 52unknown | Original sourcewinner eligible | Third-party benchmark |
18 models ranked by the best winner-eligible current-version score when available, otherwise the best publisher-artifact score; highest first. Observed Oct 19, 2026. The model you came from is highlighted.
Methodology and Coverage
Primary Evidence
Publisher Artifacts
Questions
LMArena Agent Arena FAQs
What does LMArena Agent Arena measure?+
Outcome-based measurements from real LMArena agent sessions. Model Markets classifies it as an agentic benchmark and preserves the publisher's agent-2026-10-16-a64ca3b12b6e release as a distinct comparison cohort.
How are models ranked on LMArena Agent Arena?+
Models are ordered by outcome score in score, with higher scores ranked first. Each model appears once; a verified winner-eligible current-version result takes precedence over a publisher-artifact-only result, and missing scores are not estimated.
Which model currently leads LMArena Agent Arena?+
Claude Fable 5.1 leads the current verified table with 14.03 score on agent-2026-10-16-a64ca3b12b6e. This is a benchmark-specific result, not a universal model-quality claim.
How many models have a published LMArena Agent Arena score?+
18 catalog models are published from 52 source rows. 34 source identities remain quarantined rather than guessed.
Can LMArena Agent Arena scores be compared with other benchmarks?+
Raw scores should be compared only within the same benchmark version, metric, and protocol. The Intelligence ranking normalizes its three pinned publisher inputs to a frozen reference range, requires at least one source, and labels one- and two-source rows provisional instead of estimating missing inputs. View aggregate rankings →
Why might a model be missing from LMArena Agent Arena?+
A model remains absent when the publisher has no current result, the source model identity is unresolved, the evaluation protocol is incompatible, or the evidence cannot be verified. Model Markets does not substitute a provider claim or infer a score from a related model.
How current is the LMArena Agent Arena leaderboard?+
The current Model Markets snapshot was observed Oct 19, 2026 from LMArena artifacts. The exact publisher source and each retained result artifact are linked on this page.