MME-VideoOCR

Name: MME-VideoOCR Benchmark Results
Creator: NTU
License: https://creativecommons.org/licenses/by/4.0/

NTU

1,464 videos with 2,000 QA pairs across 25 tasks. Tests OCR capabilities in video content.

Models6

Papers6

Metrics1

SOTA History

Coming Soon

Visual timeline of state-of-the-art progression over time will appear here.

Overall accuracy across all video OCR tasks

Higher is better

Rank	Model	Code	Score	Paper / Source
1	gemini-25-pro 1,464 videos, 2,000 QA pairs, 25 tasks	-	73.7%	AlphaXiv
2	qwen25-vl-72b	HF	69%	AlphaXiv
3	internvl3-78b	-	67.2%	AlphaXiv
4	gpt-4o	-	66.4%	AlphaXiv
5	gemini-15-pro	-	64.9%	AlphaXiv
6	qwen25-vl-32b	HF	61%	AlphaXiv