Open ASR · AMI-Cleaned WER ↗English short-form; normalization unversioned
9.32% WERFermionResearch/Phonon-2
FermionResearchDiscovered Oct 17, 2026
FermionResearch/Phonon-2
0.6B parameters · 1 language · Encoder: FastConformer (low-bit) · Decoder: TDT
Measured performance
All leaderboards ↗Benchmark results
Word error rate · lower is better
Open ASR · Earnings22-Cleaned-AA-chunked WER ↗English short-form; normalization unversioned
6.96% WERFermionResearch/Phonon-2
Open ASR · Gigaspeech-Cleaned WER ↗English short-form; normalization unversioned
8.35% WERFermionResearch/Phonon-2
Open ASR · LS Clean WER ↗English short-form; normalization unversioned
1.73% WERFermionResearch/Phonon-2
Open ASR · LS Other WER ↗English short-form; normalization unversioned
3.92% WERFermionResearch/Phonon-2
Open ASR · SPGISpeech WER ↗English short-form; normalization unversioned
3.7% WERFermionResearch/Phonon-2
Open ASR · URGENT2024 WER ↗English short-form; normalization unversioned
14.9% WERFermionResearch/Phonon-2
Open ASR · Voice Arena Monsoon WER ↗English short-form; normalization unversioned
4.91% WERFermionResearch/Phonon-2
Open ASR · Voxpopuli-AA-Cleaned WER ↗English short-form; normalization unversioned
2.46% WERFermionResearch/Phonon-2
Access & cost
0 quotesPrices across providers
No listed provider price
Model specifications
- Developer
- FermionResearch
- Modality
- Audio
- Context
- Not reported
- Maximum output
- Not reported
- Weights
- Not reported
- First discovered
- Oct 17, 2026