Africa's AI moment has arrived — this 1.5B model built from scratch for 12 African languages beats Google, Meta, and Alibaba while being 8x smaller
MORENA is a 1.5B African language model that beats larger rivals on benchmarks while requiring substantially less computing power.
Vambo AI has unveiled MORENA, a groundbreaking language model designed specifically for 12 African languages, including Nigerian Pidgin, Igbo, Yoruba, Hausa, Kiswahili, ChiShona, isiZulu, isiXhosa, Kinyarwanda, Setswana, Afrikaans, and isiNdebele, along with English, French, and code. This 1.5 billion parameter model outperforms major rivals such as Google, Meta, Alibaba, and HuggingFace on African language modeling and translation tasks while being significantly smaller in size.
MORENA utilizes fewer tokens to represent African text than its competitors, boasting superior performance against Google's Gemma 3 and Llama 3.2 models. The developers claim that the smaller model size contributes to a cost advantage, processing 58 times less than Lugha-Llama-8B for each byte of data. Additionally, MORENA's instruction-focused release enables conversational applications, translation, and integration with other AI tools.
Written by urgent.news from TechRadar's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
This story
This is one outlet's version. Read the fullest account.
