Speech Reasoning on MMSU S→T only
43.2AccuracyMinMo
Evaluation Results
| Method | Links | |
|---|---|---|
| MinMoType=AR2026.01 | 43.2 | |
| Phi-4-MultimodalType=AR2026.01 | 42.2 | |
| DIFFUSPEECHType=Diff.2026.01 | 39 | |
| Qwen2-AudioType=AR2026.01 | 35.7 | |
| Llama-Omni2Type=AR2026.01 | 33.4 | |
| DiFFAType=Diff.2026.01 | 29.6 | |
| MoshiType=AR2026.01 | 24 | |
| SpiritLMType=AR2026.01 | 14.5 | |
| SpeechGPTType=AR2026.01 | 9.3 |