A project by Ghana NLP to test the performance of Large Language Models (LLMs) on Ghanaian languages.
Nsanku is an ongoing community-driven initiative that evaluates how well various open-source language models perform when working with Ghanaian languages. This project aims to benchmark model capabilities and identify areas for improvement in multilingual AI systems.
As AI engineers work to bring Ghanaian languages into large language models, it’s essential to have reliable evidence on how existing models perform. Nsanku provides insights that help developers identify which open-source models are most suitable for building upon, and which languages currently have stronger or weaker support. This understanding enables realistic decisions about where to focus efforts and resources to advance multilingual AI in Ghana and across Africa.
We're currently at 75 sentences evaluated for 43 languages across 12 models thanks to our awesome contributors. Watch this video to learn how to contribute.
We welcome contributions from the community! To contribute:
- Run the evaluation using our Google Colab notebook
- Share your results with us
- We'll include your findings in our collective results
Get started with the evaluation notebook:
We are running evaluations of these models:
deepseek-v3.1gemma-2-9b-itgemma-2-27b-itgpt-oss-120bkimi-k2-instruct-0905llama-3.1-405b-instructllama-3.3-70b-instructllama-4-maverick-17b-128e-instructmistral-medium-3-instructqwen3-235b-a22bqwq-32bseed-oss-36b-instruct
The project currently evaluates 43 Ghanaian languages:
| Language | Language | Language | Language |
|---|---|---|---|
| Abron | Gikyode | Dangme | Siwu |
| Anyin | Avatime | Bisa | Bimoba |
| Southern Birifor | Tuwuli | Ntcham | Buli |
| Anufo | Dagbani | Southern Dagaare | Ewe |
| Fante | Ga | Gonja | Farefare |
| Hanga | Konni | Kusaal | Lelemi |
| Sekpele | Mampruli | Deg | Nawuri |
| Chumburung | Nkonya | Delo | Nyagbo |
| Nzema | Esahie | Paasaal | Tumulung Sisaala |
| Selee | Tafi | Tampulma | Twi |
| Vagla | Konkomba | Kasem |
We recognize our contributors:
| Name | Sentences | LLM Queries |
|---|---|---|
| Onesimus Addo Appiah | 35 | 90,300 |
| Mich-Seth Owusu | 15 | 38,700 |
| Jonathan Asiamah | 5 | 2,580 |
| Elias Dzobo | 5 | 2,580 |
| Kelvin Newman | 5 | 2,580 |
| Edmund O. Benefo | 5 | 2,580 |
| Gerhardt Datsomor | 5 | 2,580 |
| John Ayernor | 5 | 2,580 |
Your name could be here! Contribute to the project and we'll add you to our list of contributors, and you'll also get access to on of our curated datasets for Ghanaian languages
For questions or comments, please email [email protected].
To submit your contributions, send them to [email protected].
This is an open community project. We welcome researchers, developers, and language enthusiasts to participate and help advance NLP for Ghanaian languages.