Through an API

Open and closed models answer the same suite-standard 1.0 through their APIs, ranked together. 1 ranked model.

#ModelPass rate
1DeepSeek V4.1 Flash 1017Open weights100%

Model leaderboard

68 models in multi-turn ranked by LMArena score · updated Nov 11, 2026

#ModelScore
68Olmo-3.1-32B-ThinkAi21252
67gpt-oss-20bOpenAI1268
66granite-4.2-3bIBM Granite1288
65Olmo-3-32B-ThinkAi21290
64granite-4.2-8bIBM Granite1298
63Qwen3-30B-A3BQwen1307
62QwQ-32BQwen1313
61NVIDIA-Nemotron-3-Nano-30B-A3B-BF16NVIDIA1319
60Trinity-Large-ThinkingArcee AI1321
59Ling-flash-2.0inclusionAI1323
58NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16NVIDIA1324
57Qwen3-32BQwen1331
56GLM-4.5VZ.ai1333
55gpt-oss-120bOpenAI1340
54granite-4.2-30bIBM Granite1340
53GLM-4.7-FlashZ.ai1341
52Qwen3-Next-80B-A3B-ThinkingQwen1346
51INTELLECT-3Prime Intellect1350
50GLM-4.6VZ.ai1358
49MiniMax-M2MiniMax1361
Show all 68 models

Score by size

Plot against
Best for its size

Scores: LMArena, CC BY 4.0. Measured on served versions, not local builds. A model with several served variants shows its best.