Text-to-Audio Instruction Following on T2ABench
24Count Accuracy (Cnt-acc)AudioX-Base
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| AudioX-Base2026.06 | 24 | 52.8 | 17.4 | 75 | |
| Audio-Oscar2026.06 | 22.6 | 69.8 | 20.2 | — | |
| AudioX-Turbo2026.06 | 21.8 | 55.4 | 18.8 | 74.8 | |
| AudioX2026.06 | 12.4 | 23.6 | 28.2 | — | |
| Stable Audio Open2026.06 | 9.8 | 6 | 21.8 | — | |
| Stable Audio Open2026.06 | 9.8 | 6 | 21.8 | 31.2 | |
| AudioLDM-22026.06 | 7.4 | 1.2 | 13.4 | — | |
| AudioLDM-22026.06 | 7.4 | 1.2 | 13.4 | 20.1 | |
| AudioGen2026.06 | 5.4 | 6 | 18.4 | — | |
| AudioGen2026.06 | 5.4 | 6 | 18.4 | 24.4 | |
| MMAudio2026.06 | 4.8 | 2.4 | 21.4 | — | |
| MMAudio2026.06 | 4.8 | 2.4 | 21.4 | 26.6 | |
| Tango 22026.06 | 4.6 | 10.2 | 18.8 | — | |
| Tango 22026.06 | 4.6 | 10.2 | 18.8 | 25.2 | |
| AudioLDM2026.06 | 4 | 3.4 | 11.6 | — | |
| Make-An-Audio22026.06 | 4 | 19.8 | 18.8 | — | |
| AudioLDM2026.06 | 4 | 3.4 | 11.6 | 18.6 | |
| Make-An-Audio 22026.06 | 4 | 19.8 | 18.8 | 32.4 |