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

Macedonian · 6 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 whisper-large-v3most accurate 15.0914.29–15.89 2.1× 0.6× 1.55 GB Apache-2.0
2 whisper-large-v3-turbo 17.8517.07–18.66 3.4× 0.8× 0.83 GB Apache-2.0
3 Qwen3-ASR-1.7Boverall pickfast pick 18.2217.43–19.04 3.8× 2.0× 2.04 GB Apache-2.0
4 whisper-medium 24.7523.74–25.86 4.3× 1.1× 0.77 GB Apache-2.0
5 Qwen3-ASR-0.6B 35.0934.14–36.11 8.0× 4.3× 0.79 GB Apache-2.0
6 whisper-small 41.5340.36–42.67 12.1× 3.4× 0.25 GB Apache-2.0