Text-Audio Reasoning on MMAR & MMAU Hardest 1.0 (L5)
38Pass@1 AccuracyOmni-AutoThink
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Omni-AutoThinkModel Category=Omni models2025.12 | 38 | 71 | |
| Qwen3-OmniModel Category=Omni models2025.12 | 36 | 0 | |
| Qwen-AudioModel Category=Expert models2025.12 | 26 | 0 | |
| Qwen2.5-OmniModel Category=Omni models2025.12 | 6 | 0 |