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 in creative writing ranked by LMArena score · updated Nov 11, 2026

#ModelScore
1GLM-5.2Z.ai1459
2GLM-5.3Z.ai1459
3GLM-5.1Z.ai1453
4GLM-5.3-FlashZ.ai1442
5DeepSeek-V4.1-FlashDeepSeek1438
6MiMo-V2.5-ProXiaomi MiMo1437
7GLM-5Z.ai1437
8Kimi-K2.6Moonshot AI1433
9Kimi-K2.5Moonshot AI1422
10gemma-4-31B-itGoogle1415
11GLM-4.6Z.ai1412
12Hy3Tencent1405
13GLM-4.7Z.ai1402
14Qwen3.5-397B-A17BQwen1402
15gemma-4-26B-A4B-itGoogle1402
16NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16NVIDIA1399
17InklingThinking Machines1395
18MiniMax-M3MiniMax1395
19GLM-4.5Z.ai1395
20MiMo-V2.5Xiaomi MiMo1391
21Beam-1.0-501B-A23Breflection1388
22Mistral-Large-3-675B-Instruct-2512Mistral AI1386
23Qwen3-235B-A22B-Thinking-2507Qwen1384
24NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF16NVIDIA1377
25Qwen3.8-27BQwen1377
26MiMo-V2-FlashXiaomi MiMo1376
27Qwen3-235B-A22B-Instruct-2507Qwen1375
28Mistral-Medium-3.5-128BMistral AI1374
29Qwen3.5-122B-A10BQwen1368
30Qwen3-VL-235B-A22B-InstructQwen1365
31Qwen3.5-27BQwen1362
32Qwen3-235B-A22BQwen1357
33Step-3.5-FlashStepFun1357
34MiniMax-M2.7MiniMax1354
35Inkling-2-SmallThinking Machines1348
36Qwen3.5-35B-A3BQwen1347
37GLM-4.6VZ.ai1346
38LongCat-Flash-ChatMeituan LongCat1345
39Qwen3-VL-235B-A22B-ThinkingQwen1344
40GLM-4.5-AirZ.ai1343
41Mistral-Large-4-1T-A52BMistral AI1343
42Muse-Glimmer-30Bmeta-models1341
43Hy3-previewTencent1340
44Inkling-SmallThinking Machines1337
45Qwen3-Next-80B-A3B-InstructQwen1333
46Qwen3-Coder-480B-A35B-InstructQwen1333
47MiniMax-M2.5MiniMax1330
48Trinity-Large-PreviewArcee AI1319
49Qwen3-30B-A3B-Instruct-2507Qwen1317
50INTELLECT-3Prime Intellect1317
51NVIDIA-Nemotron-3-Super-120B-A12B-BF16NVIDIA1315
52Qwen3-Next-80B-A3B-ThinkingQwen1315
53Trinity-Large-ThinkingArcee AI1308
54Qwen3-32BQwen1297
55GLM-4.7-FlashZ.ai1296
56GLM-4.5VZ.ai1295
57Ling-flash-2.0inclusionAI1287
58QwQ-32BQwen1287
59MiniMax-M2MiniMax1287
60granite-4.2-30bIBM Granite1285
61gpt-oss-120bOpenAI1274
62Qwen3-30B-A3BQwen1271
63NVIDIA-Nemotron-3-Nano-30B-A3B-BF16NVIDIA1270
64Llama-4-Maverick-17B-128E-InstructMeta Llama1266
65Olmo-3.1-32B-InstructAi21266
66Llama-3.1-405B-InstructMeta Llama1262
67NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16NVIDIA1259
68granite-4.1-8bIBM Granite1251
69Llama-3.3-70B-InstructMeta Llama1250
70Llama-4-Scout-17B-16E-InstructMeta Llama1249
71granite-4.2-8bIBM Granite1235
72Llama-3.1-70B-InstructMeta Llama1232
73Qwen2.5-72B-InstructQwen1221
74granite-4.0-h-smallIBM Granite1210
75granite-4.2-3bIBM Granite1204
76gpt-oss-20bOpenAI1202
77Qwen2.5-Coder-32B-InstructQwen1174
78Llama-3.1-8B-InstructMeta Llama1154
79Mixtral-8x22B-Instruct-v0.1Mistral AI1141
80QwQ-32B-PreviewQwen1134
81granite-3.1-8b-instructIBM Granite1129
82granite-3.1-2b-instructIBM Granite1116
83Llama-3.2-3B-InstructMeta Llama1094
84granite-3.0-8b-instructIBM Granite1071
85granite-3.0-2b-instructIBM Granite1046
86Llama-3.2-1B-InstructMeta Llama1033
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.