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

86 models ranked by LMArena score · updated Nov 11, 2026

#ModelScore
86Llama-3.2-1B-InstructMeta Llama1055
85granite-3.0-2b-instructIBM Granite1080
84granite-3.0-8b-instructIBM Granite1096
83Llama-3.2-3B-InstructMeta Llama1110
82granite-3.1-2b-instructIBM Granite1127
81granite-3.1-8b-instructIBM Granite1150
80QwQ-32B-PreviewQwen1162
79Mixtral-8x22B-Instruct-v0.1Mistral AI1162
78Llama-3.1-8B-InstructMeta Llama1187
77Qwen2.5-Coder-32B-InstructQwen1230
76granite-4.0-h-smallIBM Granite1241
75Llama-3.1-70B-InstructMeta Llama1261
74Qwen2.5-72B-InstructQwen1269
73Llama-3.3-70B-InstructMeta Llama1274
72Llama-4-Scout-17B-16E-InstructMeta Llama1279
71Llama-3.1-405B-InstructMeta Llama1284
70Llama-4-Maverick-17B-128E-InstructMeta Llama1287
69gpt-oss-20bOpenAI1288
68granite-4.1-8bIBM Granite1291
67granite-4.2-3bIBM Granite1292
66Olmo-3.1-32B-InstructAi21311
65Qwen3-30B-A3BQwen1317
64granite-4.2-8bIBM Granite1319
63QwQ-32BQwen1329
62NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16NVIDIA1329
61GLM-4.5VZ.ai1333
60Trinity-Large-PreviewArcee AI1339
59Trinity-Large-ThinkingArcee AI1340
58Qwen3-32BQwen1340
57MiniMax-M2MiniMax1340
56NVIDIA-Nemotron-3-Nano-30B-A3B-BF16NVIDIA1348
55GLM-4.7-FlashZ.ai1351
54INTELLECT-3Prime Intellect1356
53Qwen3-Coder-480B-A35B-InstructQwen1357
52MiniMax-M2.5MiniMax1359
51Ling-flash-2.0inclusionAI1361
50granite-4.2-30bIBM Granite1362
49gpt-oss-120bOpenAI1366
48Qwen3-Next-80B-A3B-ThinkingQwen1368
47GLM-4.6VZ.ai1377
46NVIDIA-Nemotron-3-Super-120B-A12B-BF16NVIDIA1377
45GLM-4.5-AirZ.ai1384
44Qwen3-30B-A3B-Instruct-2507Qwen1384
43Muse-Glimmer-30Bmeta-models1389
42Qwen3-235B-A22BQwen1394
41Qwen3.5-35B-A3BQwen1395
40Qwen3-VL-235B-A22B-ThinkingQwen1401
39Hy3-previewTencent1401
38Step-3.5-FlashStepFun1402
37Inkling-2-SmallThinking Machines1403
36MiniMax-M2.7MiniMax1405
35NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF16NVIDIA1407
34Qwen3.5-27BQwen1409
33MiMo-V2-FlashXiaomi MiMo1411
32Mistral-Large-4-1T-A52BMistral AI1411
31Inkling-SmallThinking Machines1414
30Qwen3-235B-A22B-Thinking-2507Qwen1415
29Qwen3-Next-80B-A3B-InstructQwen1417
28Qwen3.5-122B-A10BQwen1417
27Qwen3-235B-A22B-Instruct-2507Qwen1419
26Qwen3-VL-235B-A22B-InstructQwen1420
25Mistral-Medium-3.5-128BMistral AI1421
24LongCat-Flash-ChatMeituan LongCat1423
23Beam-1.0-501B-A23Breflection1424
22MiMo-V2.5Xiaomi MiMo1427
21MiniMax-M3MiniMax1428
20Mistral-Large-3-675B-Instruct-2512Mistral AI1429
19GLM-4.5Z.ai1430
18gemma-4-26B-A4B-itGoogle1434
17GLM-4.7Z.ai1435
16Qwen3.8-27BQwen1437
15Qwen3.5-397B-A17BQwen1438
14Hy3Tencent1439
13GLM-4.6Z.ai1440
12gemma-4-31B-itGoogle1443
11InklingThinking Machines1443
10Kimi-K2.5Moonshot AI1445
9NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16NVIDIA1445
8GLM-5Z.ai1446
7Kimi-K2.6Moonshot AI1455
6GLM-5.1Z.ai1462
5DeepSeek-V4.1-FlashDeepSeek1464
4MiMo-V2.5-ProXiaomi MiMo1464
3GLM-5.2Z.ai1469
2GLM-5.3-FlashZ.ai1472
1GLM-5.3Z.ai1475
Show the top 20

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.