Meta's Llama models represent a family of open-source generative AI, offering developers customisation options. Unlike closed models, Llama 4 allows downloads and modifications through cloud partners like Nvidia and AWS. The Llama 4 series includes Scout, Maverick, and the upcoming Behemoth, each designed for specific tasks.
Released in April 2025, Llama 4 features models like Scout and Maverick, incorporating multimodal capabilities and a mixture-of-experts architecture. Scout excels in data analysis with a 10 million token context window, while Maverick balances speed and reasoning for coding assistance using a 1 million token window. Meta's open approach aims to challenge the dominance of closed AI models, providing flexibility for enterprise deployments and startup innovation. Llama models support over 200 languages and can perform tasks such as coding and data summarisation.
Meta uses Llama to power Meta AI across WhatsApp, Instagram and Facebook Messenger in 40 countries. Fine-tuned versions operate in over 200 countries, creating a feedback loop for model improvement. While Llama's coding performance may lag behind some closed models, its open architecture and cloud partnerships offer advantages for enterprises seeking AI customisation.
