Omni-R1: Do You Really Need Audio to Fine-Tune Your Audio LLM?
Open paper
Benchmark
Model
Rank
Results
audio-understanding-reasoning-on-mmau
Omni-R1 (FT on Qwen2.5-Omni, 8.2B)
–
Accuracy (Avg, Test-mini): 77.00
Rank counts only results with a code link.