← Arena leaderboards

Steerability

Arena · 50 models

RankModelScore
1Gemini 4 Carbon (High)0.1884
2Claude Opus 5.5 (Max)0.1853
3Claude Fable 5.1 (Max)0.1514
4GPT-6.1 Terra (Max)0.1407
5GPT-6 Astra (Max)0.1387
6GPT-6.1 Sol (Max)0.1346
7Gemini 4 Argon (High)0.1339
8Claude Sonnet 5.5 (Max)0.1127
9Muse Wonder Full (Max)0.1089
10GPT-6 Sol (Max)0.0974
11Claude Fable 5 (High)0.0888
12GPT-5.6 Sol (xHigh)0.0641
13Kimi K3.1 Pro (Max)0.0633
14Claude Opus 5 (Max)0.0531
15GPT-6 Luna (Max)0.0421
16Claude Sonnet 5 (High)0.0412
17Muse Spark 1.3 (Max)0.037
18Claude Haiku 5.5 (High)0.0346
19Muse Wonder (Max)0.0309
20DeepSeek-V4.1-Flash-20261017 (Max)0.0251
21MiMo-V2.6-Flash0.0247
22GPT-5.6 Terra (xHigh)0.0082
23MiMo-V2.6-Pro0.0075
24GPT-6 Luna (Medium)0.0055
25Kimi K3 (Max)0.0036
26Grok 4.6 (xHigh)0.0022
27Hy4 preview0.0018
28Gemini 3.8 Flash (High)-0.0008
29Qwen3.8-Max-0.0015
30Mistral Large 4-0.0049
31GLM-5.2 (Max)-0.0052
32Step 5 Preview-0.0063
33GLM-5.3 (Max)-0.0085
34DeepSeek-V4-Pro (High, 20260813)-0.0087
35Grok 4.7 (xHigh)-0.0106
36Hy3-0.024
37Qwen3.8-Flash-Next-0.0289
38GPT-6 Luna (Low)-0.0292
39GLM-5.3-Flash-0.0413
40MiniMax-M3.1-Flash-0.0469
41Qwen3.8-27B-0.0473
42Gemini 3.7 Flash (High)-0.0576
43Gemini 3.1 Pro Preview-0.0709
44Beam 1.0 501B A23B-0.0728
45MiniMax-M3-0.1048
46Qwen3.7-Plus-0.1112
47Muse Spark 1.2 (xHigh)-0.1129
48Qwen3.7-Max-0.1159
49Spark-X2.6-0.1213
50Muse Spark 1.1-0.1382