← Arena leaderboards

Text · Style Control · Instruction Following

Arena · 50 models

RankModelScore
1Gemini 4 Carbon (High)1,524.6434
2Gemini 4 Argon (High)1,524.5324
3Claude Opus 5.5 (High)1,522.1671
4Claude Opus 4.6 (High)1,510.7638
5Claude Fable 5 (High)1,508.464
6Claude Opus 4.7 (High)1,502.2105
7Claude Fable 5.1 (Max)1,500.1792
8Muse Wonder (Max)1,497.9291
9Claude Opus 4.61,497.6751
10Claude Opus 5 (High)1,496.3652
11Claude Opus 5 (Max)1,494.5498
12Claude Opus 4.71,492.6496
13Claude Opus 4.8 (High)1,491.5694
14Gemini 3.8 Flash (High)1,491.3541
15GPT-5.6 Sol (xHigh)1,490.6805
16Kimi K3 (Max)1,489.9376
17Gemini 3.7 Flash (High)1,487.0431
18Claude Sonnet 5.5 (xHigh)1,486.4102
19GPT-6.1 Sol (Max)1,485.7147
20Claude Opus 4.5 (High, 32K, 20251101)1,483.5554
21Muse Spark 1.3 (Max)1,483.2242
22Gemini 3.1 Pro Preview1,480.2329
23GLM-5.3 (Max)1,479.1513
24Claude Opus 4.81,478.075
25GPT-5.5 (High)1,477.6173
26DeepSeek-V4.1-Flash-20260910 (Max)1,476.5007
27GPT-5.51,474.9359
28Qwen3.8-Max1,474.8235
29Muse Spark 1.2 (xHigh)1,474.6739
30Claude Sonnet 4.61,474.2749
31GLM-5.3-Flash1,474.2627
32Claude Opus 4.5 (20251101)1,474.2564
33Gemini 3.6 Flash (High)1,472.7914
34Gemini 3 Pro1,472.3539
35Muse Spark 1.11,472.3165
36GPT-6 Astra (Max)1,472.2463
37MiMo-V2.6-Pro1,472.0717
38GPT-5.4 (High)1,470.9085
39GLM-5.2 (Max)1,469.0063
40MiMo-V2.5-Pro1,467.4466
41Qwen3.5-Max-Preview1,466.8777
42Claude Sonnet 5 (High)1,466.5385
43Gemini 3.5 Flash (Medium)1,466.3042
44Gemini 3.5 Flash (High)1,465.2722
45Qwen3.7-Max-Preview1,465.1731
46Muse Spark1,464.4056
47Claude Sonnet 4.5 (High, 32K, 20250929)1,463.7922
48GPT-5.6 Terra (xHigh)1,462.5134
49Claude Sonnet 4.5 (20250929)1,461.9059
50GLM-5.11,461.808