OCRBench v2

OCRBench v2 evaluates large multimodal models on bilingual visual text localization and reasoning tasks.

44rows
averageprimary metric
2026-05-06sampled

Metadata

Metrics

Average, Recognition, Referring, Spotting, Extraction, Parsing, Calculation, Understanding, Reasoning

Latest Results

Rows preserve source table rank separately for English and Chinese tasks.

Rank Subject Average Model Match Provenance Sampled
1 KDL FrontierπŸ₯‡ 68.10 β€” Imported 2026-05-06
2 NVIDIA Nemotron 3 Nano Omni 30B A3BπŸ₯ˆ 65.80 Nemotron 3 Nano Omni
nvidia-nemotron-3-nano-omni-30b-a3b-reasoning
Imported 2026-05-06
3 Gemini 3 Pro PreviewπŸ₯‰ 63.40 Gemini 3
google-gemini-3
Imported 2026-05-06
4 Seed1.6-vision 62.20 β€” Imported 2026-05-06
5 TeleMM-2.0 61.80 β€” Imported 2026-05-06
6 Qwen3-Omni-30B-A3B-Instruct 61.30 β€” Imported 2026-05-06
7 Nemotron Nano V2 VL 61.20 β€” Imported 2026-05-06
8 Gemini-2.5-Pro 59.30 Gemini 2.5 Pro
google-gemini-2.5-pro
Imported 2026-05-06
9 GLM-4.6V-Flash 59 β€” Imported 2026-05-06
10 Qwen3.5-9B 58.70 Qwen3.5-9B
qwen-qwen3.5-9b
Imported 2026-05-06
11 Llama-3.1-Nemotron-Nano-VL-8B-V1 56.40 β€” Imported 2026-05-06
12 GPT5-2025-08-07 55.50 GPT-5
openai-gpt-5
Imported 2026-05-06
13 Ovis2.5-8B 54.10 β€” Imported 2026-05-06
14 Gemini1.5-Pro 51.60 β€” Imported 2026-05-06
15 GPT-5.2 50.50 GPT-5.2
openai-gpt-5.2
Imported 2026-05-06
16 Claude Opus 4.6 48.40 Claude Opus 4.6
anthropic-claude-opus-4.6
Imported 2026-05-06
17 MiniCPM-o-4.5 44.90 β€” Imported 2026-05-06
18 Molmo2-8B 44.70 β€” Imported 2026-05-06
19 dots.mocr 43.30 β€” Imported 2026-05-06
19 Ministral-3-14B 43.30 β€” Imported 2026-05-06
21 Youtu-VL-4B-Instruct 42.30 β€” Imported 2026-05-06
22 Phi-4-multimodal-instruct 38.10 β€” Imported 2026-05-06
1 TeleMM-2.0πŸ₯‡ 66.20 β€” Imported 2026-05-06
2 Qwen3.5-9BπŸ₯ˆ 64.10 Qwen3.5-9B
qwen-qwen3.5-9b
Imported 2026-05-06
3 Gemini 3 Pro PreviewπŸ₯‰ 63.80 Gemini 3
google-gemini-3
Imported 2026-05-06
4 Gemini-2.5-Pro 62.20 Gemini 2.5 Pro
google-gemini-2.5-pro
Imported 2026-05-06
5 MiniCPM-o-4.5 61.50 β€” Imported 2026-05-06
6 KDL Frontier 61.20 β€” Imported 2026-05-06
7 Seed1.6-vision 60.50 β€” Imported 2026-05-06
8 Qwen3-Omni-30B-A3B-Instruct 60 β€” Imported 2026-05-06
9 Claude Opus 4.6 59.80 Claude Opus 4.6
anthropic-claude-opus-4.6
Imported 2026-05-06
10 GLM-4.6V-Flash 59.50 β€” Imported 2026-05-06
11 NVIDIA Nemotron 3 Nano Omni 30B A3B 59 Nemotron 3 Nano Omni
nvidia-nemotron-3-nano-omni-30b-a3b-reasoning
Imported 2026-05-06
12 MiniCPM-V-4.5-8B 58.80 β€” Imported 2026-05-06
13 SAIL-VL2-8B 57.60 β€” Imported 2026-05-06
14 Ovis2-8B 56 β€” Imported 2026-05-06
15 WeThink-Qwen2.5VL-7B 55.80 β€” Imported 2026-05-06
16 Kimi-VL-A3B-16B 54.10 β€” Imported 2026-05-06
17 Youtu-VL-4B-Instruct 54.10 β€” Imported 2026-05-06
18 GPT-5.2 52.60 GPT-5.2
openai-gpt-5.2
Imported 2026-05-06
19 dots.mocr 50.60 β€” Imported 2026-05-06
20 Ministral-3-14B 44 β€” Imported 2026-05-06
21 Phi-4-multimodal-instruct 37.30 β€” Imported 2026-05-06
22 Molmo2-8B 21.10 β€” Imported 2026-05-06