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

28 models in chinese ranked by LMArena score · updated Nov 11, 2026

#ModelScore
1GLM-5.3-FlashZ.ai1528
2Kimi-K3Moonshot AI1527
3Kimi-K2.6Moonshot AI1521
4Kimi-K3.1-ProMoonshot AI1516
5DeepSeek-V4.1-FlashDeepSeek1505
6Qwen3.5-397B-A17BQwen1498
7gemma-4-26B-A4B-itGoogle1495
8InklingThinking Machines1494
9Kimi-K2.5Moonshot AI1493
10Qwen3.8-27BQwen1493
11Qwen3.5-27BQwen1479
12gemma-4-31B-itGoogle1476
13MiMo-V2.5Xiaomi MiMo1470
14MiniMax-M3MiniMax1465
15Qwen3.5-122B-A10BQwen1463
16Inkling-SmallThinking Machines1462
17Qwen3-VL-235B-A22B-InstructQwen1462
18Mistral-Large-4-1T-A52BMistral AI1457
19Qwen3.5-35B-A3BQwen1457
20NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF16NVIDIA1451
21Qwen3-VL-235B-A22B-ThinkingQwen1442
22GLM-4.6VZ.ai1424
23Muse-Glimmer-30Bmeta-models1410
24gemma-3-27b-itGoogle1346
25GLM-4.5VZ.ai1337
26gemma-3n-E4B-itGoogle1309
27Llama-4-Maverick-17B-128E-InstructMeta Llama1278
28Llama-4-Scout-17B-16E-InstructMeta Llama1255
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.