Text-Vision Reasoning on Text-Vision Benchmark L5 1.0 (test)
54Pass@1 AccuracyQwen3-Omni
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Qwen3-OmniModel Type=Omni models2025.12 | 54 | 0 | |
| Qwen2.5-VLModel Type=Expert models2025.12 | 32 | 40 | |
| Omni-AutoThinkModel Type=Omni models2025.12 | 32 | 64 | |
| Qwen2.5-OmniModel Type=Omni models2025.12 | 25 | 0 | |
| R-4BModel Type=Expert models2025.12 | 20 | 56 |