InkubaLM has just seen a major size reduction of 75%, thanks to the Buzuzu-Mavi Challenge. This global AI competition, organized by Lelapa AI and Zindi, aimed to compress the first multilingual Small Language Model (SLM) designed for African languages.
The challenge attracted 490 participants from 61 countries, highlighting Africa’s rising talent in AI. Yvan Carré from Cameroon clinched first place by utilizing adapter heads, quantization, and knowledge distillation. Stefan Strydom from South Africa took second, trimming the model down to 40 million parameters. The AI_Buzz team from Niger and Nigeria rounded out the top three with their efficient student model.
This achievement is crucial for Africa, where approximately 33% of the population has regular internet access and 70% uses entry-level smartphones. Lean models like InkubaLM can function on devices with limited connectivity, benefiting sectors such as education and agriculture. 
CEO of Lelapa AI, Pelonomi Moiloa, stated:
“This challenge isn’t simply about technical progress, it reflects our mission to build AI that’s inclusive and grounded in African realities. When African talent is trusted with meaningful challenges, the results are transformative.”
Celina Lee, CEO of Zindi, added:
“We’re proud to show the world that more can be done with less. These solutions showcase what African innovators can achieve in the language model space.”
Moving forward, promising models from the challenge will enhance future versions of InkubaLM, which remains open-source for the global tech community. Lelapa AI and Zindi are inviting further collaboration to continue advancing African AI.
