Text-Vision Reasoning on Text-Vision Benchmark L2 1.0 (test)
0.65Pass@1 AccuracyQwen2.5-Omni
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Qwen2.5-OmniModel Type=Omni models2025.12 | 0.65 | 0 | |
| Qwen3-OmniModel Type=Omni models2025.12 | 0.64 | 0 | |
| Qwen2.5-VLModel Type=Expert models2025.12 | 0.62 | 0.38 | |
| Omni-AutoThinkModel Type=Omni models2025.12 | 0.61 | 0.67 | |
| R-4BModel Type=Expert models2025.12 | 0.16 | 0.57 |