Model card
NaturalSpeech 3.
Microsoftproprietary~500M paramsFactorized codec + non-AR diffusion1 current SOTA
Ju et al. arXiv 2403.03100.
§ 01 · Benchmarks
Every benchmark NaturalSpeech 3 has a recorded score for.
| # | Benchmark | Area · Task | Metric | Value | Rank | Date | Source |
|---|---|---|---|---|---|---|---|
| 01 | LibriTTS test-clean (Zero-Shot TTS) | Speech · Voice Cloning | wer | 1.8% | #1 | — | source ↗ |
Rank column shows this model’s position vs all other models scored on the same benchmark + metric (competitors after the slash). #1 in red means current SOTA. Sorted by rank, then newest result.
§ 02 · Strengths by area
Where NaturalSpeech 3 actually performs.
§ 04 · Related models
Other Microsoft models scored on Codesota.
§ 05 · Sources & freshness
Where these numbers come from.
editorial
1
result
0 of 1 rows marked verified.