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 non-english ranked by LMArena score · updated Nov 11, 2026

#ModelScore
68Olmo-3.1-32B-ThinkAi21209
67Olmo-3-32B-ThinkAi21253
66granite-4.2-3bIBM Granite1268
65gpt-oss-20bOpenAI1268
64Qwen3-30B-A3BQwen1295
63NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16NVIDIA1297
62granite-4.2-8bIBM Granite1300
61GLM-4.5VZ.ai1303
60QwQ-32BQwen1305
59MiniMax-M2MiniMax1313
58Qwen3-32BQwen1317
57NVIDIA-Nemotron-3-Nano-30B-A3B-BF16NVIDIA1320
56Trinity-Large-ThinkingArcee AI1326
55INTELLECT-3Prime Intellect1328
54GLM-4.7-FlashZ.ai1330
53Ling-flash-2.0inclusionAI1335
52MiniMax-M2.5MiniMax1338
51granite-4.2-30bIBM Granite1340
50Qwen3-Next-80B-A3B-ThinkingQwen1343
49gpt-oss-120bOpenAI1350
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.