Text-to-speech benchmarks
Measured resultsRaspberry Pi 5 · 16 GB RAMAll performance measurements are for this device.
Configuration
Overview
Capability & performance
Benchmark breakdown
Compare models
Word error rate ↓ · processing time ↓Listen
Unedited speech samplesListen for pronunciation, missing words, pauses and voice quality. Transcription scores do not measure how natural a voice sounds.
Test details
Reading the results
Benchmark progress
Researched models
Models we plan to test. Results appear after their benchmarks finish.
Results table · 12 configurations
Intelligibility averages word error rates from Qwen ASR and Whisper transcribing generated speech. English and German stay separate. This does not score voice naturalness; original audio samples are available above. A Pi 5 processing factor below 1 means faster than playback. A dash means no matching measurement is available.
| Model | Overall English WER · % ↓ | Overall German WER · % ↓ | Pi 5 processing factor ↓ | Peak memory · MiB ↓ |
|---|---|---|---|---|
| Kokoro 82M v1 | 3.3 | — | 1.011 | 830 |
| Piper Lessac Medium | 4.8 | — | 0.123 | 359 |
| Piper Thorsten Medium | — | 9.8 | — | — |
| Kitten Nano 0.8 | 3.3 | — | 0.241 | 315 |
| Kitten Micro 0.8 | 3.2 | — | 0.541 | 334 |
| Kitten Mini 0.8 | 3.5 | — | 1.08 | 510 |
| Ming 3.0 TTS 0.5B | 3.4 | — | 50.078 | 8,080 |
| Supertonic 3 | 3.7 | 11 | — | — |
| Pocket TTS 18.2M | 7.4 | — | — | — |
| Pocket TTS 6.9M | 6.7 | — | — | — |
| Pocket TTS 1.9M | 9.7 | — | — | — |
| Phonon 31M English | 3.5 | — | — | — |