Automatic Speech Recognition on WenetSpeech Chuan easy
10.36CERFireRedASR2S LLM
Evaluation results
| Method | Links | |
|---|---|---|
| FireRedASR2S LLMModel Size=8B+, Inference Mode=Offline2026.04 | 10.36 | |
| NIM4-ASRModel Size=2.3B, Inference Mode=Offline2026.04 | 10.51 | |
| OursModel Size=2.3B2026.04 | 10.94 | |
| Qwen3-ASR 1.7BModel Size=2.0B, Inference Mode=Offline2026.04 | 11.18 | |
| NIM4-ASRModel Size=2.3B, Inference Mode=Stream2026.04 | 11.22 | |
| Doubao-ASR2026.01 | 11.4 | |
| Qwen3-ASR-1.7BModel Size=2.0B (↓)2026.04 | 11.4 | |
| Qwen3-ASR-1.7BParameters=1.7B2026.01 | 11.99 | |
| FireRedASR-LLMModel Size=8B+ (↑)2026.04 | 12.14 | |
| Fun-ASR-nanoModel Size=0.8B (↓)2026.04 | 12.69 | |
| Fun-ASR NanoModel Size=0.8B, Inference Mode=Offline2026.04 | 13.21 | |
| Qwen3-ASR-0.6BParameters=0.6B2026.01 | 13.92 | |
| Step-Audio2-miniModel Size=8B+ (↑)2026.04 | 13.99 | |
| Step-Audio2 MiniModel Size=8B+, Inference Mode=Offline2026.04 | 13.99 | |
| Qwen3-Omni-InstModel Size=30B-A3B (↑)2026.04 | 14.13 | |
| Qwen3-Omni InstructModel Size=30B-A3B, Inference Mode=Offline2026.04 | 14.13 | |
| Whisper-large-v32026.01 | 14.35 | |
| GLM-ASR-nanoModel Size=1.5B (↓)2026.04 | 20.95 | |
| GLM-ASR NanoModel Size=1.5B, Inference Mode=Offline2026.04 | 20.95 | |
| GPT-4o-Transcribe2026.01 | 34.81 | |
| Gemini-2.5-Pro2026.01 | 43.79 |