best model by language

Benchmarked on the FLEURS test split at Q8_0. Error is WER — lower is better. Speed is × realtime on Handy target hardware.

accuracy vs speed

Vietnamese · 9 models · left is slower, down is better · log speed axis

AMD Ryzen 4750U GPU (Vulkan)
pareto frontier (no model is both faster and more accurate) other models
1 Qwen3-ASR-1.7Bmost accurate 6.155.68–6.61 3.8× 2.0× 2.04 GB Apache-2.0
2 cohere-transcribe-03-2026!overall pickfast pick 7.396.80–7.98 8.0× 3.0× 2.41 GB Apache-2.0
3 Fun-ASR-MLT-Nano-2512! 8.327.70–8.90 9.0× 4.5× 0.83 GB FunASR Model Open Source License Agreement v1.1
4 whisper-large-v3 8.748.12–9.46 2.1× 0.6× 1.55 GB Apache-2.0
5 Qwen3-ASR-0.6B 9.328.67–9.93 8.0× 4.3× 0.79 GB Apache-2.0
6 whisper-large-v3-turbo 9.488.85–10.17 3.4× 0.8× 0.83 GB Apache-2.0
7 whisper-medium 13.7412.94–14.63 4.3× 1.1× 0.77 GB Apache-2.0
8 nemotron-3.5-asr-streaming-0.6b 13.9613.09–14.83 14.5× 7.5× 0.70 GB OpenMDW-1.1
9 whisper-small 22.4721.57–23.47 12.1× 3.4× 0.25 GB Apache-2.0