Optical Character Recognition On Ocrbench V2 English
Metrics
Accuracy
Results
Performance results of various models on this benchmark
Model Name | Accuracy | Paper Title | Repository |
|---|---|---|---|
| Qwen3-Omni-30B-A3B-Instruct | 61.3 | Qwen3-Omni Technical Report | - |
| GPT5-2025-08-07 | 55.5 | - | - |
| Gemini1.5-Pro | 51.6 | - | - |
| Ovis2.5-8B | 54.1 | Ovis2.5 Technical Report | - |
| Llama-3.1-Nemotron-Nano-VL-8B-V1 | 56.4 | - | - |
| Gemini-2.5-Pro | 69.3 | - | - |
| GPT-4o | 47.6 | GPT-4o System Card | - |
| MiniCPM-V-4.5-8B | 48.4 | MiniCPM-V 4.5: Cooking Efficient MLLMs via Architecture, Data, and Training Recipe | - |
| Seed1.6-vision | 62.2 | - | - |
| SAIL-VL2-8B | 49.3 | SAIL-VL2 Technical Report | - |
0 of 10 row(s) selected.