Skip to content

Benchmark

Arena Agent Tool Hallucination

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
3.08

Score distribution

-0.002
Mean
0.008
Median
0.022
Std dev
-0.035
10th percentile
0.012
90th percentile
-0.0923
Min

This benchmark has a weight of 3.08 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
1Kimi K3.1 ProMax· LMArena data0.012
2GLM 5.3 Flash· LMArena data0.012
3GPT-6.1 SolMax· LMArena data0.012
4GPT-6.1 SolMedium· LMArena data0.012
5GPT-6.1 TerraMedium· LMArena data0.012
6GPT-6.2 AstraMax· LMArena data0.012
7GPT-6 LunaMedium· LMArena data0.012
8GPT-6.2 SolMax· LMArena data0.012
9GPT-6 AstraMax· LMArena data0.012
10Grok 4.6xHigh· LMArena data0.012
11GPT-6 LunaLow· LMArena data0.012
12GPT-6.1 TerraMax· LMArena data0.012
13Claude Fable 5.2Max· LMArena data0.012
14GLM 5.3Max· LMArena data0.012
15GPT-5.6 SolxHigh· LMArena data0.012
16Kimi K3Max· LMArena data0.012
17GPT-6 SolMax· LMArena data0.012
18Grok 4.7xHigh· LMArena data0.012
19GPT-6 LunaMax· LMArena data0.012
20GLM 5.2Max· LMArena data0.012
21GPT-5.6 TerraxHigh· LMArena data0.0119
22Qwen4 MaxxHigh· LMArena data0.0118
23Muse Spark 1.1· LMArena data0.0117
24Muse Spark 1.3Max· LMArena data0.0104
25Muse Spark 1.2xHigh· LMArena data0.0103
26Claude Opus 5Max· LMArena data0.0102
27Gemini 4 NeonHigh· LMArena data0.00987
28Claude Fable 5.1Max· LMArena data0.00976
29Claude Sonnet 5.5Max· LMArena data0.00975
30Claude Opus 5.5Max· LMArena data0.0096
31Muse EmberMax· LMArena data0.00922
32Gemini 4 CarbonHigh· LMArena data0.00821
33Gemini 3.7 FlashHigh· LMArena data0.00811
34Gemini 4 ArgonHigh· LMArena data0.00789
35Gemini 4 HydrogenHigh· LMArena data0.00705
36Claude Opus 5.5High· LMArena data0.0069
37MiniMax M3.1 Flash· LMArena data0.00483
38Claude Fable 5High· LMArena data0.00451
39Claude Sonnet 5High· LMArena data0.00417
40Hy 4· LMArena data0.00403
41Muse Wonder FullMax· LMArena data0.00283
42Mistral Large 4· LMArena data0.0018
43MiMo V2.6 Pro· LMArena data0.000197
44Gemini 3.8 FlashHigh· LMArena data0.00016
45Inkling 2 Small· LMArena data0.000138
46Qwen 3.8 27B· LMArena data-0.000286
47DeepSeek V4.1 Flash 1017Max· LMArena data-0.000565
48Muse WonderMax· LMArena data-0.00153
49Hy4 preview· LMArena data-0.00207
50MiniMax M3· LMArena data-0.00418

Showing 1–50 of 68 results.

Data credits: LMArena, CC BY 4.0 (licence)