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

32 models in hard prompts ranked by LMArena score · updated Nov 11, 2026

#ModelScore
1Kimi-K3Moonshot AI1494
2GLM-5.3-FlashZ.ai1490
3DeepSeek-V4.1-FlashDeepSeek1483
4Kimi-K3.1-ProMoonshot AI1471
5Kimi-K2.6Moonshot AI1470
6Qwen3.8-27BQwen1456
7Kimi-K2.5Moonshot AI1453
8InklingThinking Machines1453
9MiMo-V2.5Xiaomi MiMo1450
10Qwen3.5-397B-A17BQwen1448
11gemma-4-31B-itGoogle1447
12MiniMax-M3MiniMax1442
13gemma-4-26B-A4B-itGoogle1439
14Mistral-Large-4-1T-A52BMistral AI1434
15Qwen3-VL-235B-A22B-InstructQwen1428
16Inkling-SmallThinking Machines1425
17Qwen3.5-122B-A10BQwen1421
18Inkling-2-SmallThinking Machines1421
19Qwen3.5-27BQwen1414
20NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF16NVIDIA1414
21Qwen3-VL-235B-A22B-ThinkingQwen1407
22Qwen3.5-35B-A3BQwen1400
23Muse-Glimmer-30Bmeta-models1395
24GLM-4.6VZ.ai1368
25gemma-3-27b-itGoogle1340
26GLM-4.5VZ.ai1334
27gemma-3-12b-itGoogle1309
28Molmo2-8BAi21287
29gemma-3n-E4B-itGoogle1284
30Llama-4-Maverick-17B-128E-InstructMeta Llama1281
31Llama-4-Scout-17B-16E-InstructMeta Llama1266
32gemma-3-4b-itGoogle1253
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.