whatmodel.sh / tts / SILMA TTS v1
SILMA TTS v1
↗ share
+ contribute
SILMA TTS v1 is an open-weights text-to-speech model from SILMA AI, released March 2026. Its published size is 0.15B parameters. Licence: apache-2.0.
Catalogue identity notes
silma-ai/silma-tts @ 226dd7a65cadf51f9a6dbe3953fc89003b3844d5 (created 2026-03-13, Apache-2.0 weights): SILMA TTS v1, 150M-parameter F5-TTS-architecture model pretrained from scratch, Arabic (MSA) + English, zero-shot cloning. Arena identity: Arabic TTS Arena model_id silma_tts_v1_small 'SILMA TTS v1 Small' (model_url https://huggingface.co/silma-ai/silma-tts; wrapper uses the silma-tts pip package). Distinct from SILMA's commercial 'SILMA TTS v1 Large' (retired in the arena) and 'SILMA TTS v2 MSA'.
Model information sources
Release date ↗ · Weights licence ↗ · Parameter count ↗
Card: 'a high-performance, 150M-parameter bilingual (Arabic/English) TTS model'.
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 SILMA TTS v1, out of 148 with a score.
| # | model | score | where |
|---|---|---|---|
| 87 |  Parler-TTS Mini v0.1 | 49±14 | local |
| 88 |  Fish Speech 1.2 | 49±14 | local |
| 89 | SILMA TTS v1 | 48±10 | local |
| 90 |  FireRedTTS (v1) | 48±3 | local |
| 91 |  Vokan | 48±14 | comparison only |
see all text-to-speech models · find the ones that fit your machine
$ what it's good at
other tests1 not in a task · 62% of its score
| test | result | vs the others in its run | counts |
|---|---|---|---|
| Arabic TTS Arena · community blind A/B rating, 2026-09-24 snapshot | 1006.7 ✎ | ▂▂··▃·█▅···▂ beats 2/12 | 62% |
$ where to run it
Other ways to run it
- Run it as published · needs ~2.65 GB · from huggingface.co/silma-ai (official)
pip install silma-tts && python -c "from silma_tts.api import SilmaTTS; silma_tts = SilmaTTS(); silma_tts.infer(ref_file='ref.wav', ref_text=None, gen_text='أنا نموذج جديد من سلمى لتحويل النص إلى كلام.', file_wave='generated_audio.wav', seed=None, speed=1)"
Official weights and documentation: huggingface.co/silma-ai ↗
$ how the score was worked out · Arabic TTS Arena (Navid AI) ·… 62%
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 |
|---|---|---|
 Arabic TTS Arena (Navid AI) · leaderboard snapshot 2026-09-24 | 1 | 62% |
| prior · pull toward the middle | 38% | |
| weighted average | = 48.0 |
$ 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