whatmodel.sh / tts / dots.tts MF 2B (MeanFlow, NFE 4)
dots.tts MF 2B (MeanFlow, NFE 4)
↗ share
+ contribute
dots.tts MF 2B (MeanFlow, NFE 4) is an open-weights text-to-speech model from dots.tts, released June 2026. Its published size is 2B parameters. Licence: apache-2.0.
Catalogue identity notes
Exact checkpoint dots-studio/dots.tts-mf @ c28105a (Apache-2.0, created 2026-06-04): 'the CFG-aware MeanFlow distillation of dots.tts-soar', recommended at NFE = 4; the report's 'MF, NFE=4' rows are this checkpoint at its recommended setting (card prints the identical rows). Not dots.tts-soar, dots.tts-base or the later dots.tts-mf-2steps / -1step repos (2026-08-10).
Model information sources
Release date ↗ · Weights licence ↗ · Parameter count ↗
Maker's rounded '2B' system figure (Table 2 Params column); HF safetensors metadata total 2,199,404,546 (BF16).
Score within the Vapi Humanness · Bradley–Terry rating, 2026-09-25 snapshot comparison set only.
$ how it compares
The text-to-speech models ranked either side of dots.tts MF 2B (MeanFlow, NFE 4), out of 148 with a score, then more from dots.tts.
| # | model | score | where | ||
|---|---|---|---|---|---|
| 7 |  Qwen-Audio-3.0-TTS-Plus | 69±2 | api $20.00/M chars | ||
| 8 |  Gemini 3.1 Flash TTS Preview | 69±2 | api $50.00/M chars | ||
| 9 | dots.tts MF 2B (MeanFlow, NFE 4) | 68±3 | local | ||
| 10 |  Raon-OpenTTS-1B | 68±2 | comparison only | ||
| 11 |  Grok TTS | 68±3 | api $15.00/M chars | ||
| more from dots.tts | |||||
| 14 | dots.tts Base (Pretrain) 2B | 67±2 | comparison only | ||
| 20 | dots.tts SOAR 2B | 65±2 | local | ||
see all text-to-speech models · find the ones that fit your machine
$ what it's good at
voice cloning▁·▁▁▁▁▁▁▇█▃▃ #5 of 37 · 38% of its score · 4 tests
Beats Gemini 3.1 Flash TTS Preview here, though they rank higher overall. Trails dots.tts SOAR 2B, LongCat-AudioDiT-3.5B here, though they rank lower overall.
| test | result | vs the others in its run | counts |
|---|---|---|---|
| Seed-TTS test-en · speaker similarity (WavLM-large SV) lab | 76.2% ✎ | ▁·▂▂▂█▆▃▅··▂ beats 23/30 | 14% |
| Seed-TTS test-zh · speaker similarity (WavLM-large SV) lab | 80.0% ✎ | ▁·▂··▂·█▄·▃▄ beats 20/25 | 12% |
| Seed-TTS test-en · Whisper large-v3 WER lab | 1.3% ✎ | ▁···▂▂▂▂█▆▄▂ beats 33/35 | 6.2% |
| Seed-TTS test-zh · Paraformer CER lab | 0.9% ✎ | ▁·▂▁▁▁▂▃▅█·▂ beats 22/30 | 4.6% |
other tests3 not in a task · 60% of its score
| test | result | vs the others in its run | counts |
|---|---|---|---|
| CV3-Eval hard-en · WER lab | 4.4% ✎ | ▂······▅▅██▃ beats 18/18 | 24% |
| CV3-Eval EN (zero-shot English) · WER lab | 4.0% ✎ | ▁·······▁▃█▇ beats 12/14 | 20% |
| CV3-Eval hard-zh · CER lab | 9.1% ✎ | ▃▅·▃▃▅▅▃▅▃·█ beats 8/15 | 15% |
$ where to run it
Other ways to run it
- Run it as published · needs ~5.16 GB, NVIDIA only · from huggingface.co/dots-studio (official)
pip install dots.tts && dots.tts --model-name-or-path dots-studio/dots.tts-mf --text "Hello, this is a zero-shot voice cloning demonstration." --prompt-audio /path/to/reference.wav --prompt-text "The exact transcript of the reference audio." --num-steps 4 --output clone.wav
Official weights and documentation: huggingface.co/dots-studio ↗
$ how the score was worked out · dots.tts Technical Report 98%
Each result implies a score on its own, after adjusting for how hard its test is and who ran it. The score is their weighted average, with a small pull toward the middle for thinly tested models. Each test's weight is its counts column above; here it's added up by who ran the tests. How it works
| source | results | share of the score |
|---|---|---|
 dots.tts Technical Report · lab-reported | 7 | 98% |
| prior · pull toward the middle | 2% | |
| weighted average | = 68.3 |
$ improve this page
Use a short GitHub form or edit the data through a pull request. Every number needs a source; here's which ones we can use.
data updated · sources and licences · json