whatmodel.sh / tts / Ghana TTS 72k
Ghana TTS 72k
↗ share
+ contribute
Ghana TTS 72k is an open-weights text-to-speech model from GhanaNLP, released June 2026. Its published size is 0.652B parameters. Licence: mit.
Catalogue identity notes
Comparison-only. GhanaNLP community's Ghanaian-language TTS ghananlpcommunity/ghana-tts-72k (revision e8a51f1, created 2026-06-28, card licence mit, library voxcpm; card metadata lists ~40 Ghanaian languages plus English). A VoxCPM v1 fine-tune per the nsanku benchmark ('architecture': 'VoxCPM v1 (0.7B)'). The repo is gated (automatic approval after sharing contact details), so the card text was not read and no run command is listed. Added so the nsanku 7-language CER run has three catalogue models (with gemini-3-1-flash-tts-preview and omnivoice); its row there is by: lab (same maker as the benchmark).
Model information sources
Release date ↗ · Weights licence ↗
652,287,296 parameters (F32) in the HF safetensors metadata of model.safetensors; the nsanku benchmark lists the architecture as 'VoxCPM v1 (0.7B)'.
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 Ghana TTS 72k, out of 148 with a score.
| # | model | score | where |
|---|---|---|---|
| 41 |  Inworld TTS-1.5 Max | 62±3 | api $10.00/M chars |
| 42 |  ZipVoice | 62±3 | local |
| 43 | Ghana TTS 72k | 62±7 | comparison only |
| 44 |  Qwen3-TTS 12Hz 0.6B Base | 62±2 | local |
| 45 |  Eleven Flash v2 | 62±3 | api $50.00/M chars |
see all text-to-speech models · find the ones that fit your machine
$ what it's good at
voice cloning▁·▁▁▁▁▁▁██▃▃ #25 of 43 * · 79% of its score · 1 test
Beats Gemini 3.1 Flash TTS Preview, MOSS-TTS v1.5 (8B), GLM-TTS here, though they rank higher overall. Trails E2 TTS Base (F5-TTS reproduction), DMOSpeech 2, Qwen3-TTS 12Hz 0.6B Base here, though they rank lower overall.
| test | result | vs the others in its run | counts |
|---|---|---|---|
| nsanku TTS Benchmark · 7 Ghanaian languages mean CER (GhanaNLP) lab | 19.6% ✎ | █··········▄ beats 1/2 | 79% |
$ where to run it
No verified API or download recorded yet · add one
Weights and documentation: huggingface.co/ghananlpcommunity ↗
$ how the score was worked out · nsanku TTS Benchmark (GhanaNL… 79%
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 |
|---|---|---|
 nsanku TTS Benchmark (GhanaNLP community Space) · lab-reported | 1 | 79% |
| prior · pull toward the middle | 21% | |
| weighted average | = 61.8 |
$ 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