Audio Temporal Grounding on SpotSound-Bench
69R1@.3SpotSound-A
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| SpotSound-AModel Category=Open-Source Models, Integration Backbone=Audio Flamingo 32026.04 | 69 | 53.3 | 52.7 | |
| SpotSound-QModel Category=Open-Source Models, Integration Backbone=Qwen2 Audio2026.04 | 62.3 | 45 | 46.6 | |
| WTATGModel Category=Non-LLM Models2026.04 | 47 | 26.7 | 32.3 | |
| Gemini-2.5-FlashModel Category=Proprietary Models2026.04 | 32.7 | 28 | 23.2 | |
| AM-DETRModel Category=Non-LLM Models2026.04 | 30 | 19.7 | 22.5 | |
| Gemini-2.5-ProModel Category=Proprietary Models2026.04 | 19.7 | 17 | 18.9 | |
| Audio Flamingo 3Model Category=Open-Source Models2026.04 | 10.7 | 3.7 | 9.1 | |
| TimeAudioModel Category=Open-Source Models2026.04 | 8.7 | 1.3 | 11 | |
| Qwen2-AudioModel Category=Open-Source Models2026.04 | 7.3 | 3.3 | 6.2 | |
| Kimi-AudioModel Category=Open-Source Models2026.04 | 2.7 | 0.7 | 2.4 |