ByteBrief
We're a portrait publication through and through. Turn your phone back and your briefing picks up right where you left it.
(We tried widescreen once. It wasn't us.)

Vambo AI released MORENA, a 1.5 billion parameter language model trained from scratch for 12 African languages plus English, French, and code. It outperforms larger models from Google, Meta, Alibaba, and HuggingFace on African language tasks while using up to 8x fewer parameters. The model scored 1.408 bpb on a key benchmark, beating rivals with over five times as many parameters. Development used over 22,000 A100 GPU hours valued at roughly $40,000, supported by UNDP, AIHub4SD, and CINECA. Smaller 0.5B and 0.2B versions are also available for speech recognition and offline use.
Tap to vote and see what everyone thinks.
Summary by ByteBrief