Audio Understanding & Reasoning on WavCaps
39.4GPTQwen2-Audio + CoAT
Evaluation Results
| Method | Links | |
|---|---|---|
| Qwen2-Audio + CoATBackbone=Qwen2-Audio, Method=CoAT2026.06 | 39.4 | |
| Qwen2.5-Omni-7B + CoATBackbone=Qwen2.5-Omni-7B, Method=CoAT2026.06 | 33.6 | |
| Audio Flamingo 3Backbone=Audio Flamingo 3, Method=Baseline2026.06 | 33.4 | |
| Qwen2-AudioBackbone=Qwen2-Audio, Method=Baseline2026.06 | 32.2 | |
| Qwen2.5-Omni-7BBackbone=Qwen2.5-Omni-7B, Method=Baseline2026.06 | 29.4 | |
| Audio Flamingo 3 + CoATBackbone=Audio Flamingo 3, Method=CoAT2026.06 | 29 |