handy models

speech-to-text benchmarks for transcribe.cpp, the engine behind Handy

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

Spanish · 27 models

AMD Ryzen 4750U GPU (Vulkan)
pareto frontier (no model is both faster and more accurate) other models
highlighted rows are recommended picks or on pareto frontier
1 Voxtral-Mini-4B-Realtime-2602 2.67 0.8× 0.6× 4.41 GB Apache-2.0
2 whisper-large-v3 2.70 2.0× 1.0× 1.55 GB Apache-2.0
3 Voxtral-Small-24B-2507 2.86 — — 24.04 GB Apache-2.0
4 parakeet-ultraoverall pickaccurate pickfast pick 3.06 23.0× 13.3× 0.69 GB CC-BY-4.0
5 canary-1b-v2! 3.10 14.1× 8.3× 1.07 GB CC-BY-4.0
6 whisper-large-v3-turbo 3.12 3.1× 1.4× 0.83 GB Apache-2.0
7 whisper-large-v2 3.30 2.0× 1.0× 1.55 GB Apache-2.0
8 Qwen3-ASR-1.7B 3.31 4.1× 2.5× 2.04 GB Apache-2.0
9 Voxtral-Mini-3B-2507 3.52 1.1× 0.6× 4.66 GB Apache-2.0
10 whisper-large 3.55 2.0× 1.0× 1.55 GB Apache-2.0
11 parakeet-tdt-0.6b-v3 3.65 25.1× 13.7× 0.69 GB CC-BY-4.0
12 orukeet 3.75 25.1× 13.6× 0.69 GB CC-BY-SA-4.0
13 whisper-medium 3.80 4.1× 1.9× 0.77 GB Apache-2.0
14 parakeet-primeline 3.85 25.0× 13.6× 0.69 GB CC-BY-4.0
15 cohere-transcribe-03-2026! 3.97 8.1× 4.3× 2.25 GB Apache-2.0
16 granite-speech-4.1-2b-nar! 4.08 4.0× 2.1× 2.33 GB Apache-2.0
17 Qwen3-ASR-0.6B 4.88 9.7× 5.3× 0.79 GB Apache-2.0
18 granite-speech-4.1-2b! 5.48 4.6× 2.8× 2.38 GB Apache-2.0
19 granite-4.0-1b-speech! 5.91 4.6× 2.8× 2.38 GB Apache-2.0
20 whisper-small 5.92 11.7× 5.9× 0.25 GB Apache-2.0
21 canary-1b! 6.06 9.8× 6.5× 1.08 GB CC-BY-NC-4.0
22 nemotron-3.5-asr-streaming-0.6b 6.30 17.1× 10.6× 0.70 GB OpenMDW-1.1
23 granite-speech-4.1-2b-plus! 6.53 4.6× 2.6× 2.18 GB Apache-2.0
24 canary-180m-flash! 6.54 33.1× 21.7× 0.20 GB CC-BY-4.0
25 canary-1b-flash! 6.73 15.5× 8.8× 0.98 GB CC-BY-4.0
26 whisper-base 11.15 29.0× 19.3× 0.08 GB Apache-2.0
27 whisper-tiny 18.95 43.1× 39.8× 0.04 GB Apache-2.0