Skip to content

Benchmark

Arena Agent Task Outcome Explicit

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
2.51

Score distribution

0
Mean
0.022
Median
0.117
Std dev
-0.146
10th percentile
0.142
90th percentile
-0.318
Min
-0.3180.225
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 2.51 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 data0.225
2Claude Opus 5.5Max· LMArena data0.192
3Gemini 4 NeonHigh· LMArena data0.192
4Claude Sonnet 5.5Max· LMArena data0.176
5Gemini 4 CarbonHigh· LMArena data0.172
6Gemini 4 HydrogenHigh· LMArena data0.157
7GPT-6 AstraMax· LMArena data0.149
8Claude Fable 5.1Max· LMArena data0.142
9Gemini 4 ArgonHigh· LMArena data0.126
10Claude Opus 5.5High· LMArena data0.126
11GPT-6.2 AstraMax· LMArena data0.124
12GPT-6.1 SolMax· LMArena data0.116
13GPT-6.2 SolMax· LMArena data0.109
14Muse Wonder FullMax· LMArena data0.104
15Claude Opus 5Max· LMArena data0.101
16Qwen4 MaxxHigh· LMArena data0.0882
17GPT-6.1 TerraMax· LMArena data0.0853
18Claude Sonnet 5.5High· LMArena data0.0706
19Hy 4· LMArena data0.0632
20GPT-6.1 SolMedium· LMArena data0.056
21Claude Fable 5High· LMArena data0.0441
22GPT-6 SolMax· LMArena data0.039
23Qwen3.8 Max· LMArena data0.0359
24Grok 4.7xHigh· LMArena data0.0351
25MiMo V2.6 Pro· LMArena data0.0348
26Hy4 preview· LMArena data0.0325
27Kimi K3Max· LMArena data0.0274
28GLM 5.3Max· LMArena data0.0267
29GPT-5.6 SolxHigh· LMArena data0.0253
30Muse Spark 1.3Max· LMArena data0.025
31MiMo V2.6 Flash· LMArena data0.0249
32Muse WonderMax· LMArena data0.0249
33Inkling 2 Small· LMArena data0.0244
34DeepSeek V4.1 Flash 1017Max· LMArena data0.0224
35Kimi K3.1 ProMax· LMArena data0.0193
36GLM 5.2Max· LMArena data0.0116
37Gemini 3.8 FlashHigh· LMArena data0.01
38GLM 5.3 Flash· LMArena data0.00634
39GPT-6.1 TerraMedium· LMArena data0.00437
40Qwen 3.8 Flash Next· LMArena data0.00254
41Claude Haiku 5.5High· LMArena data0.00228
42Claude Sonnet 5High· LMArena data-0.00919
43GPT-6 LunaMax· LMArena data-0.012
44Muse EmberMax· LMArena data-0.0199
45Step 5 Preview· LMArena data-0.0214
46Grok 4.6xHigh· LMArena data-0.0247
47MiniMax M3.1 Flash· LMArena data-0.0261
48Qwen 3.8 27B· LMArena data-0.0326
49Gemini 3.7 FlashHigh· LMArena data-0.0345
50Muse Spark 1.1· LMArena data-0.0726

Showing 1–50 of 68 results.

Data credits: LMArena, CC BY 4.0 (licence)