Automatic Speech Recognition on MLS En
3.68WERGemini-2.5-Pro
Evaluation results
| Method | Links | |
|---|---|---|
| Gemini-2.5-Pro2026.01 | 3.68 | |
| Qwen3-Omni InstructModel Size=30B-A3B, Inference Mode=Offline2026.04 | 4.04 | |
| Step-Audio2 MiniModel Size=8B+, Inference Mode=Offline2026.04 | 4.37 | |
| Qwen3-ASR-1.7BParameters=1.7B2026.01 | 4.58 | |
| FireRedASR2S LLMModel Size=8B+, Inference Mode=Offline2026.04 | 4.71 | |
| NIM4-ASRModel Size=2.3B, Inference Mode=Offline2026.04 | 4.77 | |
| Whisper-large-v32026.01 | 4.87 | |
| Qwen3-ASR 1.7BModel Size=2.0B, Inference Mode=Offline2026.04 | 4.93 | |
| NIM4-ASRModel Size=2.3B, Inference Mode=Stream2026.04 | 5.04 | |
| GPT-4o-Transcribe2026.01 | 5.12 | |
| GLM-ASR NanoModel Size=1.5B, Inference Mode=Offline2026.04 | 5.32 | |
| Fast ConformerDecoder=RNNT2023.05 | 5.76 | |
| ConformerDecoder=RNNT2023.05 | 5.77 | |
| Qwen3-ASR-0.6BParameters=0.6B2026.01 | 6 | |
| Fast ConformerDecoder=CTC2023.05 | 6.42 | |
| ConformerDecoder=CTC2023.05 | 6.6 | |
| Fun-ASR NanoModel Size=0.8B, Inference Mode=Offline2026.04 | 6.8 | |
| Doubao-ASR2026.01 | 7.09 |