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

141 models in math ranked by LMArena score · updated Nov 11, 2026

#ModelScore
1GLM-5.3-FlashZ.ai1503
2GLM-5.3Z.ai1500
3Kimi-K3Moonshot AI1496
4GLM-5.2Z.ai1483
5InklingThinking Machines1481
6MiMo-V2.5-ProXiaomi MiMo1481
7DeepSeek-V4.1-FlashDeepSeek1479
8Hy3Tencent1477
9Kimi-K2.6Moonshot AI1475
10GLM-5.1Z.ai1472
11gemma-4-26B-A4B-itGoogle1470
12Kimi-K2.5Moonshot AI1469
13gemma-4-31B-itGoogle1462
14Beam-1.0-501B-A23Breflection1459
15Inkling-SmallThinking Machines1459
16NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16NVIDIA1456
17Mistral-Large-4-1T-A52BMistral AI1456
18Qwen3.8-27BQwen1455
19Qwen3.5-397B-A17BQwen1453
20DeepSeek-V4-ProDeepSeek1442
Show all 141 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.