Skip to content

Benchmark

Arena ELO Vision - English

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

Models measured
147
Scores recorded
168
Sources
1
IQ weight
1.75

Score distribution

1,193
Mean
1,230.5
Median
107.3
Std dev
1,012.8
10th percentile
1,297.8
90th percentile
836.8
Min
836.81,325.5
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.75 for chat models, measured over 147 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 Opus 5.5High· LMArena data1,325.5
2Claude Fable 5High· LMArena data1,325.3
3Claude Opus 4.6High· LMArena data1,324.2
4Claude Opus 4.7· LMArena data1,321.6
5Claude Opus 4.7High· LMArena data1,319.6
6Claude Fable 5.1Max· LMArena data1,318.6
7Claude Opus 4.6· LMArena data1,318.4
8Gemini 3.8 FlashHigh· LMArena data1,317.8
9Claude Fable 5.2Max· LMArena data1,314.8
10Claude Opus 5High· LMArena data1,310.4
11Gemini 3.7 FlashHigh· LMArena data1,308.9
12Qwen3.8 Max· LMArena data1,305.9
13Muse Spark· LMArena data1,304.3
14Gemini 3.5 FlashHigh· LMArena data1,304.3
15GLM 5.3 Flash· LMArena data1,299.8
16GPT 5.4High· LMArena data1,298.3
17Muse Spark 1.3Max· LMArena data1,298.2
18Claude Sonnet 4.6· LMArena data1,297.8
19Gemini 3.5 FlashMedium· LMArena data1,297.8
20GPT 5.5High· LMArena data1,293.8
21Doubao Seed 2.0 Pro· LMArena data1,293.6
22Claude Opus 4.8· LMArena data1,293.5
23Muse Spark 1.3xHigh· LMArena data1,292.9
24Muse Spark 1.2xHigh· LMArena data1,292.4
25Gemini 3 Pro Preview· LMArena data1,291.8
26GPT 5.5· LMArena data1,291.6
27DeepSeek V4.1 Flash 0910Max· LMArena data1,290.9
28Muse Spark 1.1· LMArena data1,290.6
29Claude Opus 4.8High· LMArena data1,290.4
30GPT 5.4· LMArena data1,289.9
31Gemini 3.6 FlashHigh· LMArena data1,288
32Kimi K2.6· LMArena data1,287
33Gemini 3.1 Pro Preview· LMArena data1,286.3
34GPT-5.6 SolxHigh· LMArena data1,284.3
35Grok 4.5· LMArena data1,284.1
36Claude Sonnet 5.5xHigh· LMArena data1,280.5
37GPT-6 AstraMax· LMArena data1,280.1
38Qwen3.7 Plus· LMArena data1,278.9
39GPT-6.1 SolMax· LMArena data1,277.1
40MiMo V2.6 Flash· LMArena data1,275.7
41Step 5 Preview· LMArena data1,275.2
42GPT-5.2 Chat20260210· LMArena data1,275.1
43Gemini 3 Flash· LMArena data1,274.8
44Kimi K2.5Thinking· LMArena data1,274.8
45Qwen 3.8 27B· LMArena data1,274.8
46Gemma 4 31B IT· LMArena data1,273.5
47Claude Sonnet 5High· LMArena data1,273.4
48MiMo V2.6 Pro· LMArena data1,270.4
49GPT-5.6 TerraxHigh· LMArena data1,269.2
50Qwen3.5 397B A17B· LMArena data1,268.7

Showing 1–50 of 168 results.

Data credits: LMArena, CC BY 4.0 (licence)