Text-Vision Reasoning on Text-Vision Benchmark L4 1.0 (test)
57Pass@1 AccuracyQwen3-Omni
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Qwen3-OmniModel Type=Omni models2025.12 | 57 | 0 | |
| Omni-AutoThinkModel Type=Omni models2025.12 | 42 | 72 | |
| Qwen2.5-VLModel Type=Expert models2025.12 | 41 | 45 | |
| Qwen2.5-OmniModel Type=Omni models2025.12 | 37 | 0 | |
| R-4BModel Type=Expert models2025.12 | 23 | 59 |