Skip to content

Benchmark

Arena Agent Steerability

Agentic benchmarks measure how well models carry out tasks that take several steps. Higher scores are better.

Models measured
62
Scores recorded
68
Sources
1
IQ weight
2.12

Score distribution

0
Mean
0.008
Median
0.118
Std dev
-0.147
10th percentile
0.14
90th percentile
-0.298
Min
-0.2980.245
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 2.12 for chat models, measured over 62 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
1Claude Fable 5.2Max· LMArena data0.245
2Gemini 4 NeonHigh· LMArena data0.212
3Claude Opus 5.5Max· LMArena data0.18
4Gemini 4 CarbonHigh· LMArena data0.179
5Gemini 4 HydrogenHigh· LMArena data0.174
6Claude Sonnet 5.5Max· LMArena data0.148
7Claude Opus 5.5High· LMArena data0.141
8GPT-6.2 AstraMax· LMArena data0.14
9GPT-6 AstraMax· LMArena data0.139
10Claude Fable 5.1Max· LMArena data0.135
11GPT-6.2 SolMax· LMArena data0.134
12GPT-6.1 TerraMax· LMArena data0.132
13Gemini 4 ArgonHigh· LMArena data0.122
14GPT-6.1 SolMedium· LMArena data0.118
15GPT-6.1 SolMax· LMArena data0.115
16GPT-6.1 TerraMedium· LMArena data0.0991
17Claude Sonnet 5.5High· LMArena data0.095
18GPT-6 SolMax· LMArena data0.0926
19Hy 4· LMArena data0.085
20Qwen4 MaxxHigh· LMArena data0.0803
21Muse Wonder FullMax· LMArena data0.0745
22Claude Fable 5High· LMArena data0.0553
23Muse EmberMax· LMArena data0.0478
24Kimi K3.1 ProMax· LMArena data0.045
25Inkling 2 Small· LMArena data0.0436
26GPT-5.6 SolxHigh· LMArena data0.038
27Claude Opus 5Max· LMArena data0.0245
28GPT-6 LunaMax· LMArena data0.024
29DeepSeek V4.1 Flash 1017Max· LMArena data0.0233
30MiMo V2.6 Flash· LMArena data0.0162
31Muse WonderMax· LMArena data0.0153
32Claude Sonnet 5High· LMArena data0.0121
33Muse Spark 1.3Max· LMArena data0.00849
34Claude Haiku 5.5High· LMArena data0.00782
35MiMo V2.6 Pro· LMArena data0.00142
36GPT-6 LunaMedium· LMArena data-0.0094
37Hy4 preview· LMArena data-0.0108
38Qwen3.8 Max· LMArena data-0.0189
39GPT-5.6 TerraxHigh· LMArena data-0.0201
40GLM 5.2Max· LMArena data-0.0217
41Kimi K3Max· LMArena data-0.0221
42Grok 4.6xHigh· LMArena data-0.0249
43Gemini 3.8 FlashHigh· LMArena data-0.0278
44Step 5 Preview· LMArena data-0.0281
45MiniMax M3.1 Flash· LMArena data-0.0304
46GLM 5.3Max· LMArena data-0.0333
47GPT-6 LunaLow· LMArena data-0.0357
48Qwen 3.8 Flash Next· LMArena data-0.0385
49Grok 4.7xHigh· LMArena data-0.0429
50Hy3 (Deep Infra)· LMArena data-0.0492

Showing 1–50 of 68 results.

Data credits: LMArena, CC BY 4.0 (licence)