Reasoning on AIME 2025 (Score)
98.3AIME 2025 ScoreAudex 30B-A3B
Evaluation Results
| Method | Links | |
|---|---|---|
| Audex 30B-A3BContext Length=1M, Tool-integrated reasoning=true2026.07 | 98.3 | |
| Audex 30B-A3BContext Length=1M, Tool-integrated reasoning=false2026.07 | 91.2 | |
| Qwen3-Omni 30B-A3B ThinkingContext Length=64K2026.07 | 73.7 | |
| Audex 2BContext Length=128K, Tool-integrated reasoning=false2026.07 | 68.9 | |
| Audex 2BContext Length=128K, Tool-integrated reasoning=true2026.07 | 68.3 | |
| Step-Audio R1.1 33BContext Length=64K2026.07 | 44.8 |