Announcement
Introducing Phonon-1
An open speech recognition model for English: a 415 MB download that runs on a laptop or a datacenter GPU and transcribes an hour of audio in about two and a half minutes.
Speech recognition · Available nowPhonon-1 Micro
The smallest Phonon model: open speech recognition for English in a 285 MB download, state of the art at its size.
- 285 MB
- Download, 331 MB on disk
- 3.002 %
- LibriSpeech test-clean
- 8 of 8
- Benchmarks ahead of Moonshine base
- Apache-2.0
- Open weights
Phonon-1 Micro is an open speech recognition model for English, the smallest build of the Phonon family. It downloads in 285 MB, runs on a laptop or a datacenter GPU, and is state of the art at its size. It was trained at 2.4 bits per weight from the start, part of the lab’s second low-bit release after Neutrino-1.
It is more accurate than Moonshine base, the nearest smaller model, on all eight benchmarks in the table below, a macro word error rate of 8.52 against 10.1. Output arrives punctuated and capitalized from the model itself.
- Phonon-1 Microthe smallest model
- 285 MB download · 331 MB on disk
- Phonon-1the default model
- 415 MB download · 455 MB on disk
- Language
- English
- Audio inputmicrophone or file
- 16 kHz
- Output
- Punctuated, capitalized text
Benchmarks
| Benchmark | Phonon-1415 MB | Phonon-1 Micro285 MB | Parakeet-0.6B 4-bit637 MB | Moonshine base248 MB | Whisper large-v3-turbo1,619 MB | Whisper small967 MB | wav2vec2-large1,262 MB | Qwen3-ASR teacher1,569 MB |
|---|---|---|---|---|---|---|---|---|
| LibriSpeech test-clean | 2.640 | 3.002 | 2.186 | 3.417 | 2.10 | 3.4† | 2.8† | 2.235 |
| LibriSpeech test-other | 5.699 | 6.511 | 3.937 | 8.262 | 4.07 | 7.6† | 6.3† | 4.618 |
| TED-LIUM | 3.421 | 3.878 | 2.829 | 5.272 | – | – | – | 2.889 |
| SPGISpeech | 4.163 | 4.858 | 4.104 | 5.731 | 2.79† | – | 13.31† | 3.074 |
| VoxPopuli | 8.394 | 9.177 | 6.345 | 10.470 | 11.22† | – | – | 7.151 |
| GigaSpeech | 11.396 | 11.882 | 9.614 | 12.114 | 8.52† | – | – | 9.321 |
| Earnings-22 | 12.571 | 14.771 | 11.190 | 17.872 | 11.07† | – | 36.28† | 11.188 |
| AMI | 13.084 | 14.094 | 12.723 | 17.790 | 15.16† | – | – | 12.560 |
| Macro, eight benchmarks | 7.67 | 8.52 | 6.62 | 10.1 | – | – | – | 6.63 |
Run it
$ pip install fermion-research$ fermion transcribe --model FermionResearch/Phonon-1-Micro recording.wav
One-shot file transcription. The same flag selects the model for fermion listen and fermion serve. The same weights run on a Mac (via MLX) or an NVIDIA GPU.
$ fermion listen --model FermionResearch/Phonon-1-Micro$ fermion serve --model FermionResearch/Phonon-1-Micro
$ docker run --rm --gpus all ghcr.io/fermionresearch/phonon-cuda:latest \transcribe --model FermionResearch/Phonon-1-Micro recording.wav
The container transcribes recordings of any length, streams over WebSocket, and serves the same endpoint as fermion serve.
Availability
Phonon-1 Micro ships as public weights on Hugging Face with a command line on PyPI. No waitlist, no gated preview.
Speech recognition that stays on the device.
Open weightsApache-2.0
License
Open weights under the Apache License 2.0. Commercial use, modification, fine-tuning, and redistribution are permitted, with no access request and no acceptance form. The model is a derivative of Qwen/Qwen3-ASR-0.6B, itself Apache-2.0; the command line is Apache-2.0 too.
Citation
@misc{fermionresearch2026phononmicro,
title = {Phonon-1 Micro},
author = {{Fermion Research}},
year = {2026},
url = {https://fermionresearch.com/models/phonon-1-micro/}
}