Skip to content

Benchmark

Arena ELO Vision - Chinese

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

Models measured
118
Scores recorded
135
Sources
1
IQ weight
2.18

Score distribution

1,237.1
Mean
1,277.2
Median
130.5
Std dev
1,009
10th percentile
1,368
90th percentile
889.6
Min
889.61,400.6
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 2.18 for chat models, measured over 118 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 5High· LMArena data1,400.6
2Claude Fable 5.1Max· LMArena data1,394.4
3Gemini 3.8 FlashHigh· LMArena data1,391
4Gemini 3.7 FlashHigh· LMArena data1,388.3
5Claude Fable 5High· LMArena data1,377.2
6Claude Sonnet 5.5xHigh· LMArena data1,376.7
7Claude Opus 5.5High· LMArena data1,376.1
8Gemini 3.5 FlashMedium· LMArena data1,375.9
9Gemini 3.5 FlashHigh· LMArena data1,374.4
10Claude Opus 4.7· LMArena data1,372.2
11Claude Opus 4.6· LMArena data1,370.7
12Muse Spark 1.3Max· LMArena data1,370.1
13Muse Spark 1.3xHigh· LMArena data1,368.2
14Gemini 3 Pro Preview· LMArena data1,368
15Claude Opus 4.6High· LMArena data1,364.4
16Gemini 3.1 Pro Preview· LMArena data1,361.8
17GPT 5.5· LMArena data1,358.5
18GLM 5.3 Flash· LMArena data1,357.5
19Claude Opus 4.7High· LMArena data1,357.2
20GPT 5.5High· LMArena data1,356.6
21GPT 5.4High· LMArena data1,356.3
22Gemini 3.6 FlashHigh· LMArena data1,354.5
23Qwen3.8 Max· LMArena data1,353.7
24Muse Spark 1.1· LMArena data1,348.4
25Kimi K2.6· LMArena data1,347.7
26Muse Spark· LMArena data1,346.3
27Claude Opus 4.8· LMArena data1,345.3
28DeepSeek V4.1 Flash 0910Max· LMArena data1,344.3
29Claude Opus 4.8High· LMArena data1,341
30MiMo V2.6 Flash· LMArena data1,338.4
31Doubao Seed 2.0 Pro· LMArena data1,337.4
32Grok 4.5· LMArena data1,337
33GPT-6 AstraMax· LMArena data1,337
34GPT 5.4· LMArena data1,335.6
35Claude Sonnet 5High· LMArena data1,334.3
36Qwen3.7 Plus· LMArena data1,332.3
37GPT-6.1 SolMax· LMArena data1,332.1
38Mistral Large 4· LMArena data1,331
39Gemini 3 Flash· LMArena data1,330.8
40Gemma 4 31B IT· LMArena data1,327.4
41Kimi K2.5Instant· LMArena data1,327
42Qwen 3.8 27B· LMArena data1,327
43GPT-5.6 TerraxHigh· LMArena data1,324.9
44Kimi K2.5Thinking· LMArena data1,324.5
45MiMo V2.5· LMArena data1,322.5
46Step 5 Preview· LMArena data1,318
47Gemini 3 FlashMinimal· LMArena data1,316.6
48Qwen3.5 397B A17B· LMArena data1,314.2
49Gemma 4 26B A4B· LMArena data1,311.6
50Claude Sonnet 4.6· LMArena data1,309.9

Showing 1–50 of 135 results.

Data credits: LMArena, CC BY 4.0 (licence)