1,464 videos with 2,000 QA pairs across 25 tasks. Tests OCR capabilities in video content.
Total Accuracy is the reported evaluation metric for MME-VideoOCR. Codesota tracks published model scores on this metric so readers can compare state-of-the-art results across sources and model families.
Higher is better
Muted rows were not state of the art when published — an earlier or same-year result already scored better.
| Rank | Model | Trust | Score | Year | Links | Fix |
|---|---|---|---|---|---|---|
| 01 | gemini-25-pro | paper | 73.7 | 2025 | Source ↗ | Looks wrong? |
| 02 | Gemini 2.5 Pro | unverified | 73.7 | 2025 | Source ↗ | Looks wrong? |
| 03 | Qwen2.5-VL 72B | unverified | 69 | 2025 | Source ↗ | Looks wrong? |
| 04 | qwen25-vl-72b | paper | 69 | 2025 | Source ↗ | Looks wrong? |
| 05 | internvl3-78b | paper | 67.2 | 2025 | Source ↗ | Looks wrong? |
| 06 | gpt-4o | paper | 66.4 | 2025 | Source ↗ | Looks wrong? |
| 07 | gemini-15-pro | paper | 64.9 | 2025 | Source ↗ | Looks wrong? |
| 08 | Gemini 1.5 Pro | unverified | 64.9 | 2025 | Source ↗ | Looks wrong? |
| 09 | Qwen2.5-VL 32B | unverified | 61 | 2025 | Source ↗ | Looks wrong? |
| 10 | qwen25-vl-32b | paper | 61 | 2025 | Source ↗ | Looks wrong? |