Automatic Speech Recognition on AISHELL-2 (test)
2.71CERQwen3-ASR-1.7B
Evaluation results
| Method | Links | |
|---|---|---|
| Qwen3-ASR-1.7BParameters=1.7B2026.01 | 2.71 | |
| Doubao-ASR2026.01 | 2.85 | |
| Qwen3-ASR-0.6BParameters=0.6B2026.01 | 3.15 | |
| GPT-4o-Transcribe2026.01 | 4.24 | |
| ProposedModel type=decoder-only, Streaming=false2026.01 | 5 | |
| Whisper-large-v32026.01 | 5.06 | |
| ProposedModel type=decoder-only, Streaming=true2026.01 | 5.5 | |
| BESTOWModel type=decoder-only, Streaming=true2026.01 | 5.6 | |
| Baseline-non-streamModel type=encoder-decoder, Streaming=false2026.01 | 5.9 | |
| WeNet-U2Model type=encoder-decoder, Streaming=false2026.01 | 6.1 | |
| Baseline-streamModel type=encoder-decoder, Streaming=true2026.01 | 6.1 | |
| BTIModel type=decoder-only, Streaming=true2026.01 | 7.2 | |
| Gemini-2.5-Pro2026.01 | 11.62 |