handy models

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

model detail

Where this model stands against the best model in each language at Q8_0, its speed on an AMD Ryzen 4750U laptop, and what the card says. Add a second model to compare them head to head.

model moonshine-tiny-zh
AMD 4750U GPU42.9×
AMD 4750U CPU28.4×
Q8_0 size0.03 GB
frontier · AMD 4750U GPU1 of 1

verdict: Behind the leader's 95% CI in its one benchmarked language, Mandarin Chinese (CER). Faster than realtime on the AMD 4750U GPU (42.9×) and AMD 4750U CPU (28.4×).

license: MIT · card languages: 1 · —

standing by language · 1 benchmarked

Each bar compares this model with the best other model in that language, on a log scale: left of center it is ahead, right of center behind, and the other model is named on the row. The two error rates use the legend colors: this model first, the other model second. Faded bars are ties, where the 95% confidence intervals overlap. Sorted strongest standing first. Hover a row for full details.

moonshine-tiny-zh error best other model error tie: the 95% CIs overlap

quantization · 1 usable language at F32, F16

F32 is within noise on the one usable language: 310% of the Q8_0 file for no measurable accuracy cost. F16 is within noise on the one usable language: 167% of the Q8_0 file for no measurable accuracy cost.

F32 file310% of Q8_00.10 GB against 0.03 GB
median error change−1.3%over 1 usable language
measurably worse0 of 1above the Q8_0 95% confidence interval
worst usable languageMandarin Chinese (CER)13.88 → 13.70 (−0.18 pts)
F16 file167% of Q8_00.06 GB against 0.03 GB
median error change−1.3%over 1 usable language
measurably worse0 of 1above the Q8_0 95% confidence interval
worst usable languageMandarin Chinese (CER)13.88 → 13.70 (−0.18 pts)
measurably worse than Q8_0: above its 95% confidence interval within noise typical noise of one measurement
per-language numbers (1)
Mandarin Chinese (CER) 13.8812.83–14.93 13.70−0.1813.70−0.18 −1.3%

speed · all 73 models on the AMD 4750U GPU

× realtime, log scale. Left of the realtime marker, transcribing takes longer than the recording. Hover a dot for the model.

compare with

moonshine-base-zhsame familyQwen3-ASR-1.7Bahead of it in Mandarin ChineseBreeze-ASR-25closest coverage, 1 language in common