Skip to content

Benchmark

Arena ELO - Expert

People compare model answers and choose which they prefer. Higher scores are better.

Models measured
318
Scores recorded
371
Sources
1
IQ weight
1.30

Score distribution

1,333.9
Mean
1,371.6
Median
146.5
Std dev
1,113
10th percentile
1,491.8
90th percentile
953.7
Min
953.71,581.3
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.30 for chat models, measured over 318 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
1Gemini 4 CarbonHigh· LMArena data1,581.3
2Gemini 4 NeonHigh· LMArena data1,559.7
3Claude Opus 5.5High· LMArena data1,558
4Gemini 4 ArgonHigh· LMArena data1,555.6
5Claude Opus 5Max· LMArena data1,552.8
6Gemini 4 HydrogenHigh· LMArena data1,552.6
7Claude Opus 5High· LMArena data1,551.9
8Claude Fable 5.1Max· LMArena data1,546.4
9Claude Opus 4.6High· LMArena data1,542.9
10MiMo V2.6 Pro· LMArena data1,539.8
11Claude Opus 4.6· LMArena data1,537.7
12Claude Fable 5High· LMArena data1,534.6
13Gemini 3.8 FlashHigh· LMArena data1,526.3
14Claude Opus 4.7· LMArena data1,520.4
15Claude Opus 4.7High· LMArena data1,519.9
16GLM 5.3Max· LMArena data1,518.5
17Muse WonderMax· LMArena data1,516
18GPT-5.6 SolxHigh· LMArena data1,515.8
19Gemini 3.7 FlashHigh· LMArena data1,515.5
20Kimi K3Max· LMArena data1,515.3
21GLM 5.3 Flash· LMArena data1,513.1
22Muse Spark 1.3Max· LMArena data1,512.2
23Claude Sonnet 5.5xHigh· LMArena data1,511.3
24Claude Fable 5.2Max· LMArena data1,510.1
25Qwen3.8 Max· LMArena data1,509.7
26nova-experimental-chat26-02-10· LMArena data1,508.5
27GPT 5.4High· LMArena data1,506.7
28GPT 5.5High· LMArena data1,506.6
29GPT 5.5· LMArena data1,502.9
30Claude Opus 4.8High· LMArena data1,502.4
31GPT-6.1 SolMax· LMArena data1,502.4
32DeepSeek V4.1 Flash 0910Max· LMArena data1,502.3
33MiMo V2.5 Pro· LMArena data1,501.2
34Claude Sonnet 4.6· LMArena data1,500.4
35GPT-5.6 TerraxHigh· LMArena data1,494.5
36Gemini 3.5 FlashHigh· LMArena data1,494.3
37GPT-6 AstraMax· LMArena data1,493.1
38ERNIE 5.1· LMArena data1,491.8
39Kimi K2.6· LMArena data1,491.5
40MiMo V2.6 Flash· LMArena data1,491.3
41Claude Opus 4.8· LMArena data1,490.7
42Qwen3.7 Max Preview· LMArena data1,489
43Qwen3.5-Max-Preview· LMArena data1,488.9
44GLM 5.2Max· LMArena data1,488.7
45Gemini 3.6 FlashHigh· LMArena data1,488.6
46Kimi K3.1 ProMax· LMArena data1,488.3
47Claude Sonnet 5High· LMArena data1,488.1
48Claude Opus 4.520251101· LMArena data1,487.8
49Muse Spark 1.2xHigh· LMArena data1,486.3
50Gemini 3.5 FlashMedium· LMArena data1,486.3

Showing 1–50 of 371 results.

Data credits: LMArena, CC BY 4.0 (licence)