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

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

#ModelScore
1Kimi-K3Moonshot AI1465
2GLM-5.3Z.ai1462
3GLM-5.3-FlashZ.ai1462
4GLM-5.2Z.ai1458
5Kimi-K3.1-ProMoonshot AI1450
6DeepSeek-V4.1-FlashDeepSeek1450
7GLM-5.1Z.ai1448
8MiMo-V2.5-ProXiaomi MiMo1448
9Kimi-K2.6Moonshot AI1446
10DeepSeek-V4-ProDeepSeek1439
11InklingThinking Machines1436
12Kimi-K2.5Moonshot AI1432
13GLM-5Z.ai1431
14gemma-4-31B-itGoogle1430
15Qwen3.5-397B-A17BQwen1430
16NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16NVIDIA1428
17Hy3Tencent1426
18GLM-4.6Z.ai1426
19Qwen3.8-27BQwen1424
20gemma-4-26B-A4B-itGoogle1421
21DeepSeek-V4-FlashDeepSeek1419
22GLM-4.7Z.ai1417
23GLM-4.5Z.ai1417
24Mistral-Large-3-675B-Instruct-2512Mistral AI1414
25MiniMax-M3MiniMax1414
26DeepSeek-R1-0528DeepSeek1412
27Beam-1.0-501B-A23Breflection1412
28Qwen3-235B-A22B-Instruct-2507Qwen1409
29DeepSeek-V3.2DeepSeek1409
30DeepSeek-V3.1-TerminusDeepSeek1408
31Qwen3-Next-80B-A3B-InstructQwen1407
32Mistral-Medium-3.5-128BMistral AI1405
33Qwen3-VL-235B-A22B-InstructQwen1405
34DeepSeek-V3.2-ExpDeepSeek1404
35MiMo-V2.5Xiaomi MiMo1404
36LongCat-Flash-ChatMeituan LongCat1404
37Inkling-SmallThinking Machines1402
38Qwen3.5-122B-A10BQwen1400
39DeepSeek-V3.1DeepSeek1400
40Qwen3-235B-A22B-Thinking-2507Qwen1397
41Mistral-Large-4-1T-A52BMistral AI1397
42MiMo-V2-FlashXiaomi MiMo1392
43Inkling-2-SmallThinking Machines1390
44Hy3-previewTencent1390
45Qwen3.5-27BQwen1389
46NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF16NVIDIA1386
47Step-3.5-FlashStepFun1385
48Muse-Glimmer-30Bmeta-models1385
49Qwen3-235B-A22BQwen1384
50MiniMax-M2.7MiniMax1383
51Qwen3.5-35B-A3BQwen1378
52Qwen3-VL-235B-A22B-ThinkingQwen1377
53Qwen3-30B-A3B-Instruct-2507Qwen1372
54GLM-4.5-AirZ.ai1366
55GLM-4.6VZ.ai1359
56DeepSeek-V3-0324DeepSeek1358
57DeepSeek-R1DeepSeek1357
58NVIDIA-Nemotron-3-Super-120B-A12B-BF16NVIDIA1355
59gpt-oss-120bOpenAI1350
60Qwen3-Coder-480B-A35B-InstructQwen1345
61Qwen3-Next-80B-A3B-ThinkingQwen1343
62granite-4.2-30bIBM Granite1340
63MiniMax-M2.5MiniMax1338
64Ling-flash-2.0inclusionAI1335
65gemma-3-27b-itGoogle1334
66GLM-4.7-FlashZ.ai1330
67INTELLECT-3Prime Intellect1328
68Trinity-Large-ThinkingArcee AI1326
69NVIDIA-Nemotron-3-Nano-30B-A3B-BF16NVIDIA1320
70Trinity-Large-PreviewArcee AI1318
71gemma-3-12b-itGoogle1318
72Qwen3-32BQwen1317
73DeepSeek-V3DeepSeek1316
74Mistral-Small-3.2-24B-Instruct-2506Mistral AI1315
75MiniMax-M2MiniMax1313
76QwQ-32BQwen1305
77GLM-4.5VZ.ai1303
78granite-4.2-8bIBM Granite1300
79Ring-flash-2.0inclusionAI1299
80NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16NVIDIA1297
81Qwen3-30B-A3BQwen1295
82gemma-3n-E4B-itGoogle1285
83Olmo-3.1-32B-InstructAi21276
84Molmo2-8BAi21274
85DeepSeek-V2.5-1210DeepSeek1273
86gemma-3-4b-itGoogle1273
87Llama-4-Maverick-17B-128E-InstructMeta Llama1269
88gpt-oss-20bOpenAI1268
89granite-4.2-3bIBM Granite1268
90granite-4.1-8bIBM Granite1261
91Mistral-Small-3.1-24B-Instruct-2503Mistral AI1255
92DeepSeek-V2.5DeepSeek1253
93Olmo-3-32B-ThinkAi21253
94Llama-4-Scout-17B-16E-InstructMeta Llama1252
95Qwen2.5-72B-InstructQwen1252
96Llama-3.1-405B-InstructMeta Llama1248
97Llama-3.1-Nemotron-70B-InstructNVIDIA1245
98Llama-3.3-70B-InstructMeta Llama1236
99Llama-3.1-Tulu-3-70BAi21236
100Llama-3.1-70B-InstructMeta Llama1219
101gemma-2-27b-itGoogle1217
102granite-4.0-h-smallIBM Granite1217
103aya-expanse-32bCohere Labs1213
104Olmo-3.1-32B-ThinkAi21209
105Nemotron-4-340B-InstructNVIDIA1206
106Qwen2.5-Coder-32B-InstructQwen1205
107Mistral-Small-24B-Instruct-2501Mistral AI1198
108phi-4Microsoft1197
109gemma-2-9b-itGoogle1188
110DeepSeek-Coder-V2-InstructDeepSeek1182
111aya-expanse-8bCohere Labs1180
112Qwen2-72B-InstructQwen1176
113Llama-3.1-Tulu-3-8BAi21169
114Llama-3.1-8B-InstructMeta Llama1148
115Qwen1.5-110B-ChatQwen1142
116Meta-Llama-3-70B-InstructMeta Llama1142
117Qwen1.5-72B-ChatQwen1135
118QwQ-32B-PreviewQwen1130
119Mixtral-8x22B-Instruct-v0.1Mistral AI1128
120gemma-2-2b-itGoogle1121
121Qwen1.5-32B-ChatQwen1106
122granite-3.1-8b-instructIBM Granite1099
123Meta-Llama-3-8B-InstructMeta Llama1098
124Qwen1.5-14B-ChatQwen1095
125Phi-3-medium-4k-instructMicrosoft1094
126Mixtral-8x7B-Instruct-v0.1Mistral AI1077
127deepseek-llm-67b-chatDeepSeek1073
128granite-3.1-2b-instructIBM Granite1068
129Qwen1.5-7B-ChatQwen1058
130Phi-3-small-8k-instructMicrosoft1058
131gemma-1.1-7b-itGoogle1052
132Llama-2-70b-chat-hfMeta Llama1045
133Qwen-14B-ChatQwen1041
134granite-3.0-8b-instructIBM Granite1037
135granite-3.0-2b-instructIBM Granite1033
136Llama-2-13b-chat-hfMeta Llama1024
137Phi-3-mini-4k-instructMicrosoft1021
138Llama-3.2-3B-InstructMeta Llama1019
139SOLAR-10.7B-Instruct-v1.0Upstage1014
140Mistral-7B-Instruct-v0.2Mistral AI1012
141Phi-3-mini-128k-instructMicrosoft1000
142gemma-7b-itGoogle999
143gemma-1.1-2b-itGoogle988
144Qwen1.5-4B-ChatQwen979
145OLMo-7B-InstructAi2977
146Llama-3.2-1B-InstructMeta Llama973
147Llama-2-7b-chat-hfMeta Llama973
148gemma-2b-itGoogle958
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.