Skip to content

Benchmark

Arena ELO Vision - Ocr

OCR benchmarks measure how well models read text in images and documents. Higher scores are better.

Models measured
108
Scores recorded
124
Sources
1
IQ weight
1.83

Score distribution

1,247.8
Mean
1,256.6
Median
59.1
Std dev
1,156.8
10th percentile
1,318.7
90th percentile
1,058.5
Min
1,058.51,336.4
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.83 for chat models, measured over 108 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 5High· LMArena data1,336.4
2Gemini 3.8 FlashHigh· LMArena data1,331.3
3Claude Opus 5.5High· LMArena data1,331
4Claude Fable 5.1Max· LMArena data1,329.8
5Claude Opus 4.7· LMArena data1,327.1
6Claude Fable 5.2Max· LMArena data1,326.8
7Claude Opus 4.6· LMArena data1,326.4
8Claude Opus 5High· LMArena data1,325.3
9Claude Opus 4.6High· LMArena data1,325.1
10Gemini 3.7 FlashHigh· LMArena data1,321.2
11Claude Opus 4.7High· LMArena data1,320.7
12Qwen3.8 Max· LMArena data1,318.9
13Muse Spark 1.3Max· LMArena data1,318.7
14Gemini 3.5 FlashMedium· LMArena data1,315.6
15Muse Spark 1.2xHigh· LMArena data1,311.7
16Gemini 3.5 FlashHigh· LMArena data1,310.7
17GPT 5.4High· LMArena data1,310.1
18Gemini 3.6 FlashHigh· LMArena data1,309.7
19Gemini 3 Pro Preview· LMArena data1,309.6
20Muse Spark 1.3xHigh· LMArena data1,306.8
21Muse Spark· LMArena data1,306.6
22GPT 5.5High· LMArena data1,305.7
23GLM 5.3 Flash· LMArena data1,305.5
24Gemini 3.1 Pro Preview· LMArena data1,304.5
25GPT 5.5· LMArena data1,303.6
26Claude Opus 4.8High· LMArena data1,303.2
27Claude Sonnet 5.5xHigh· LMArena data1,299.6
28Muse Spark 1.1· LMArena data1,299.5
29GPT 5.4· LMArena data1,298.8
30Claude Opus 4.8· LMArena data1,297.6
31Claude Sonnet 4.6· LMArena data1,295.3
32Grok 4.5· LMArena data1,292.8
33Kimi K2.6· LMArena data1,291.4
34GPT-5.6 SolxHigh· LMArena data1,290.8
35Gemini 3 Flash· LMArena data1,290.2
36GPT-6 AstraMax· LMArena data1,288.8
37Qwen3.7 Plus· LMArena data1,288.2
38GPT-6.1 SolMax· LMArena data1,287.3
39Claude Sonnet 5High· LMArena data1,285.4
40Gemma 4 31B IT· LMArena data1,285.2
41Doubao Seed 2.0 Pro· LMArena data1,283.3
42DeepSeek V4.1 Flash 0910Max· LMArena data1,283.2
43Qwen 3.8 27B· LMArena data1,278.9
44Step 5 Preview· LMArena data1,278.7
45Gemini 3.5 Flash Lite· LMArena data1,278.4
46GPT-5.6 TerraxHigh· LMArena data1,278.2
47Kimi K2.5Thinking· LMArena data1,276.1
48GPT-5.2 Chat20260210· LMArena data1,274.6
49MiMo V2.6 Pro· LMArena data1,274.4
50Grok 4.6High· LMArena data1,271.9

Showing 1–50 of 124 results.

Data credits: LMArena, CC BY 4.0 (licence)