Speaker Information Modeling on AudioMarathon 1.0 (test)
72.3SD (Score)Qwen2.5-Omni-7B
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| Qwen2.5-Omni-7BParameters=7B2026.02 | 72.3 | 53.4 | 21.4 | 98 | |
| Voxtral-Mini-3B-2507Parameters=3B, Version=25072026.02 | 68 | 29.7 | 30.7 | 71 | |
| Qwen2.5-Omni-3BParameters=3B2026.02 | 67.3 | 39.6 | 29.1 | 97.2 | |
| Baichuan-Omni-1.52026.02 | 49.2 | 18.9 | 10.2 | 81.5 | |
| Audio-Flamingo-22026.02 | 45.9 | 13.1 | 20.3 | 85.1 | |
| Gemma-3n-E4B-itModel Size=4B2026.02 | 35.9 | 18.9 | 21.8 | 93 | |
| Gemma-3n-E2B-itModel Size=2B2026.02 | 35.1 | 15.2 | 12.2 | 91.6 | |
| Gemini-2.5-Flash-Lite2026.02 | 33.9 | 14.6 | 19.6 | 77.9 | |
| Audio-Flamingo-32026.02 | 33.7 | 54.3 | 40.7 | 96.2 | |
| Aero-1-Audio2026.02 | 33.7 | 32 | 17.8 | 47.5 | |
| Speech-XL2026.02 | 33.3 | 30.3 | 26.6 | 77.8 | |
| Gemini-2.5-Flash2026.02 | 33.1 | 31.9 | 34.3 | 99.3 | |
| GPT-4o-Audio (2024-10-01)Release Date=2024-10-012026.02 | 32.5 | 22.5 | 17.2 | 69.2 | |
| GPT-4o-Audio (2024-12-17)Release Date=2024-12-172026.02 | 30.8 | 21.8 | 19.9 | 73.1 | |
| Phi-4-Multimodal2026.02 | 26.4 | 27.3 | 26.6 | 91.1 |