OCRBench

OCRBench evaluates OCR capabilities of large multimodal models across text recognition, scene-text VQA, document VQA, KIE, and HMER.

29rows
final_scoreprimary metric
2026-05-06sampled

Metadata

Metrics

Text recognition, Scene text-centric VQA, Document-oriented VQA, Key information extraction, Handwritten mathematical expression recognition, Final score

Latest Results

Rows ranked by Final Score.

Rank Subject Final score Model Match Provenance Sampled
1 Minicpm-V 2.6 852 Imported 2026-05-06
2 granite-vision-3.3-2b 824 Imported 2026-05-06
3 MiniMonkey 806 Imported 2026-05-06
4 H2OVL-Mississippi-2B 782 Imported 2026-05-06
5 InternVL2-1B 779 Imported 2026-05-06
6 InternVL2-4B 776 Imported 2026-05-06
7 InternVL2-2B 768 Imported 2026-05-06
8 H2OVL-Mississippi-0.8B 751 Imported 2026-05-06
9 Qwen-VL-Max 723 Qwen VL Max
qwen-qwen-vl-max
Imported 2026-05-06
10 Qwen-VL-Plus 694 Qwen VL Plus
qwen-qwen-vl-plus
Imported 2026-05-06
11 Gemini 659 Imported 2026-05-06
12 GPT4V 645 GPT-4
openai-gpt-4
Imported 2026-05-06
13 MiniCPM-V-2 605 Imported 2026-05-06
14 mPLUG-DocOwl1.5 599 Imported 2026-05-06
15 TextMonkey 561 Imported 2026-05-06
16 InternVL-Chat-Chinese 517 Imported 2026-05-06
17 Monkey 514 Imported 2026-05-06
18 InternLM-XComposer2 511 Imported 2026-05-06
19 QwenVL 506 Imported 2026-05-06
20 mPLUG-Owl2 366 Imported 2026-05-06
21 LLaVAR 346 Imported 2026-05-06
22 LLaVA1.5-13B 331 Imported 2026-05-06
23 InternLM-XComposer 303 Imported 2026-05-06
24 LLaVA1.5-7B 297 Imported 2026-05-06
25 mPLUG-Owl 297 Imported 2026-05-06
26 BLIVA 291 Imported 2026-05-06
27 InstructBLIP 276 Imported 2026-05-06
28 BLIP2-6.7B 235 Imported 2026-05-06
29 MiniGPT4V2 157 Imported 2026-05-06