Skip to content
ModelCap

Benchmark leaderboard

LMArena Agent 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.

Tracked models with a result
56
Entries on the source board
68
Open-weight models listed
11

LMArena Agent standings (56 models)

Sorted by published score, one row per model. Source: LMArena.

LMArena Agent leaderboard: every tracked model's published score, source rank, ModelCap Index position, price and context
#ModelScoreConfigurationSource rankModelCap IndexOutput $/1MContextWeightsPublished
1Claude Fable 5.2Anthropic27.4%24.5%–30.4%Max#1 / 68#6$40.001MNo linked weights18 November 2026
2Claude Opus 5.5Anthropic18.6%16.6%–20.7%Max+1 more evaluated#2 / 68#4$20.001MNo linked weights18 November 2026
3Gemini 4 NeonGoogle18.4%15.9%–20.9%High#3 / 68#2$15.001MNo linked weights18 November 2026
4Gemini 4 CarbonGoogle16.1%15.4%–16.8%High#4 / 68#1$10.001MNo linked weights18 November 2026
5Claude Sonnet 5.5Anthropic15.6%12.7%–18.5%Max+1 more evaluated#5 / 68#24$10.001MNo linked weights18 November 2026
6GPT-6.2 AstraOpenAI15.2%12.5%–18.0%Max#6 / 68#22$40.001MNo linked weights18 November 2026
7GPT-6 AstraOpenAI14.6%12.5%–16.6%Max#7 / 68—$50.001MNo linked weights18 November 2026
8Claude Fable 5.1Anthropic13.7%11.6%–15.7%Max#9 / 68—$50.001MNo linked weights18 November 2026
9Gemini 4 HydrogenGoogle13.5%11.8%–15.3%High#10 / 68#9$7.501MNo linked weights18 November 2026
10GPT-6.2 SolOpenAI11.3%8.9%–13.7%Max#11 / 68#29$10.001MNo linked weights18 November 2026
11GPT-6.1 SolOpenAI11.3%10.2%–12.4%Max+1 more evaluated#12 / 68—$10.001MNo linked weights18 November 2026
12GPT-6.1 TerraOpenAI10.5%9.3%–11.7%Max+1 more evaluated#13 / 68#37$5.001MNo linked weights18 November 2026
13Gemini 4 ArgonGoogle10.3%9.4%–11.2%High#14 / 68#3$10.001MNo linked weights18 November 2026
14Qwen4 MaxQwen7.3%5.9%–8.6%X-High#18 / 68#36$4.271MNo linked weights18 November 2026
15GPT-6 SolOpenAI6.5%4.9%–8.0%Max#19 / 68—$10.001MNo linked weights18 November 2026
16Claude Opus 5Anthropic6.3%5.0%–7.6%Max#20 / 68—$25.001MNo linked weights18 November 2026
17Claude Fable 5Anthropic5.9%4.6%–7.1%High#22 / 68—$50.001MNo linked weights18 November 2026
18Kimi K3.1 ProMoonshot AI3.9%2.8%–5.0%Max#24 / 68#11$8.801MRestricted license18 November 2026
19GPT-5.6 SolOpenAI3.7%2.6%–4.8%X-High#25 / 68—$20.001MNo linked weights18 November 2026
20Inkling 2 SmallThinking Machines1.5%0.5%–2.5%Default#26 / 68#98$1.001MOpen weights18 November 2026
21Gemini 3.8 FlashGoogle1.5%0.6%–2.4%High#27 / 68#8$4.131MNo linked weights18 November 2026
22Grok 4.7xAI1.5%0.0%–3.0%X-High#28 / 68#59$6.00500KNo linked weights18 November 2026
23MiMo-V2.6-ProXiaomi1.4%0.5%–2.3%Default#29 / 68#20$0.871MOpen weights18 November 2026
24Muse Spark 1.3Meta1.1%0.2%–2.0%Max#30 / 68#7$4.251MNo linked weights18 November 2026
25Claude Sonnet 5Anthropic1.0%-0.2%–2.2%High#31 / 68—$10.001MNo linked weights18 November 2026
26Muse EmberMeta1.0%0.0%–1.9%Max#32 / 68#63$0.501MNo linked weights18 November 2026
27GPT-6 LunaOpenAI0.6%0.2%–1.1%Max+2 more evaluated#34 / 68#58$0.501MNo linked weights18 November 2026
28Muse WonderMeta0.6%-0.7%–1.9%Max#35 / 68#5$20.001MNo linked weights18 November 2026
29GLM 5.3Z.ai0.0%-0.9%–0.8%Max#36 / 68#19$1.401MRestricted license18 November 2026
30Kimi K3Moonshot AI-0.1%-0.8%–0.5%Max#37 / 68#17$16.001MRestricted license18 November 2026
31GLM 5.2Z.ai-0.4%-1.2%–0.4%Max#38 / 68—$6.801MOpen weights18 November 2026
32Hy4 previewTencent-0.4%-1.2%–0.3%Default#39 / 68#45$2.501MOpen weights18 November 2026
33Grok 4.6xAI-0.6%-1.9%–0.7%X-High#40 / 68—$6.00500KNo linked weights18 November 2026
34Claude Haiku 5.5Anthropic-0.6%-1.6%–0.4%High#41 / 68#93$0.501MNo linked weights18 November 2026
35Qwen3.8 Max (0803)Qwen-0.9%-1.6%–-0.2%Default#42 / 68—$6.001MRestricted license18 November 2026
36MiMo-V2.6-FlashXiaomi-1.8%-2.9%–-0.7%Default#43 / 68#46$0.281MOpen weights18 November 2026
37GPT-5.6 TerraOpenAI-2.7%-3.5%–-1.8%X-High#44 / 68—$12.001MNo linked weights18 November 2026
38Gemini 3.7 FlashGoogle-3.1%-4.0%–-2.1%High#46 / 68—$3.751MNo linked weights18 November 2026
39MiniMax M3.1 FlashMiniMax-3.7%-5.1%–-2.2%Default#47 / 68#33$0.601MRestricted license18 November 2026
40GLM 5.3 FlashZ.ai-3.9%-4.6%–-3.1%Default#48 / 68#26$0.501MOpen weights18 November 2026
41Step 5 PreviewStepFun-5.1%-6.6%–-3.7%Default#50 / 68#49$2.701MRestricted license18 November 2026
42Qwen3.8 27BQwen-5.5%-6.4%–-4.6%Default#51 / 68#68$2.201MOpen weights18 November 2026
43Gemini 3.1 Pro PreviewGoogle-5.7%-7.4%–-3.9%Default#52 / 68#16$12.001MNo linked weights18 November 2026
44Mistral Large 4Mistral AI-7.1%-9.2%–-5.1%Default#54 / 68#96$2.101MRestricted license18 November 2026
45Qwen3.7 MaxQwen-8.3%-9.6%–-6.9%Default#55 / 68—$4.421MNo linked weights18 November 2026
46Beam 1.0 501B A23BReflection-8.3%-11.2%–-5.4%Default#56 / 68#67$1.001MOpen weights18 November 2026
47Hy3Tencent-8.6%-10.2%–-7.1%Default#57 / 68—$0.528262KOpen weights18 November 2026
48Qwen3.7 PlusQwen-9.0%-10.6%–-7.4%Default#58 / 68#43$1.281MNo linked weights18 November 2026
49Muse Spark 1.1Meta-9.0%-9.7%–-8.4%Default#59 / 68—$4.251MNo linked weights18 November 2026
50Muse Spark 1.2Meta-9.1%-10.0%–-8.2%X-High#60 / 68—$4.251MNo linked weights18 November 2026
51MiniMax M3MiniMax-10.9%-12.1%–-9.7%Default#61 / 68#66$1.201MRestricted license18 November 2026
52Mistral Medium 3.5Mistral AI-14.5%-16.2%–-12.7%Default#63 / 68#87$7.50262KRestricted license18 November 2026
53Nemotron 3 UltraNVIDIA-16.8%-20.2%–-13.4%Default#64 / 68#261$2.20262KRestricted license18 November 2026
54InklingThinking Machines-18.5%-20.0%–-16.9%Default#66 / 68#62$4.05524KOpen weights18 November 2026
55Inkling SmallThinking Machines-18.5%-20.6%–-16.5%Default#67 / 68—$1.20524KOpen weights18 November 2026
56Solar Pro 4Upstage-23.9%-26.1%–-21.6%Default#68 / 68#145$0.60524KNo linked weights18 November 2026

How ModelCap uses LMArena Agent

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 Agent leaderboard: common questions

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.

Which AI model leads LMArena Agent right now?

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.

How many models are ranked on the LMArena Agent leaderboard here?

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.

What is the best open-weight model on LMArena Agent?

Inkling 2 Small from Thinking Machines is the highest-scoring model with openly downloadable weights on this board, at 1.5%.

Which model offers the best value on LMArena Agent?

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%.

How is LMArena Agent scored?

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.

Does LMArena Agent decide the ModelCap Index rank?

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.

How recent are the LMArena Agent results?

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.

Explore further