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
51Gemini 3 FlashMinimal· LMArena data1,271.7
52Gemini 2.5 Pro· LMArena data1,270.3
53Qwen3.5 397B A17B· LMArena data1,270.2
54MiMo V2.6 Flash· LMArena data1,269.6
55Gemma 4 26B A4B· LMArena data1,267.2
56Grok 4.20Reasoning, 0309, beta· LMArena data1,266.3
57GPT-5.6 LunaxHigh· LMArena data1,263.3
58GPT-6.1 TerraMax· LMArena data1,262.1
59Grok 4.20 Multi-Agent0309, beta· LMArena data1,261.5
60Mistral Large 4· LMArena data1,260.9
61Kimi K2.5Instant· LMArena data1,258.3
62GPT 5.2High· LMArena data1,256.6
63GPT-5.5 Instant· LMArena data1,256.3
64MiMo V2.5· LMArena data1,256.2
65MiniMax M3· LMArena data1,254.9
66Gemini 2.5 Flash Preview09-2025· LMArena data1,254.7
67GPT 5.1High· LMArena data1,253.5
68GPT-5.4 miniHigh· LMArena data1,253.1
69Qwen3 VL 235B· LMArena data1,252
70GPT-6 SolMax· LMArena data1,250.8
71Qwen3.5 122B-A10B· LMArena data1,250.2
72ERNIE-5.0-Preview1220· LMArena data1,249.1
73ChatGPT-4o Latest20250326· LMArena data1,248.8
74Gemini 3.1 Flash Lite· LMArena data1,245.9
75GPT-5 Chat· LMArena data1,245
76Qwen3.5 27B· LMArena data1,244.2
77Grok 4.7xHigh· LMArena data1,244.1
78Inkling Small· LMArena data1,242.5
79GPT 5.1· LMArena data1,242.4
80MiMo V2 Omni· LMArena data1,238.4
81Gemini 2.5 Flash· LMArena data1,237.5
82Qwen-VL Max2025-08-13· LMArena data1,235.5
83Grok 4.3· LMArena data1,233.8
84GPT 5.2· LMArena data1,232.8
85Mistral Medium 3.5· LMArena data1,232.4
86Mistral Large 3· LMArena data1,229.1
87NVIDIA Nemotron 3.5 Super VL 120BNVFP4· LMArena data1,222.3
88GPT-6 LunaMax· LMArena data1,219.8
89GPT 4.12025-04-14· LMArena data1,218.1
90OpenAI o32025-04-16· LMArena data1,217.9
91GPT 5High· LMArena data1,214.7
92Qwen3 VL 235Bthinking· LMArena data1,212.1
93GPT 5.4 NanoHigh· LMArena data1,203.3
94Gemini 2.5 Flash-Lite PreviewNon-Thinking, 09-2025· LMArena data1,201.7
95GPT-5 miniHigh· LMArena data1,198.7
96o4 Mini2025-04-16· LMArena data1,196.7
97Claude Opus 4Thinking 16K, 20250514· LMArena data1,196.7
98Claude Sonnet 4Thinking 32K, 20250514· LMArena data1,196.5
99Grok 4.1 FastReasoning· LMArena data1,194.8
100Grok 40709· LMArena data1,194.6

Showing 51–100 of 124 results.

Data credits: LMArena, CC BY 4.0 (licence)