AI updates
Llama 4 Scout and Maverick: Meta moves to mixture-of-experts
Meta's Llama 4 introduces two natively multimodal MoE models with 17B active parameters each: Scout (~109B total, 16 experts) and Maverick (~400B total, 128 experts), trained on up to 40T tokens.
- Both models activate 17B parameters per token; Scout uses 16 experts and Maverick 128.
- Early fusion makes them natively multimodal for text and images.