Meta integrates Llama 3.1 405B into WhatsApp, doubles down on open-source

Meta is leaning further into its open-source Llama ecosystem, integrating the Llama 3.1 405B model into WhatsApp to deliver free generative AI capabilities to its global user base, while continuing to position the 'Llama 4 herd' as natively multimodal. The strategy is explicit: commoditize foundational models so that enterprises build proprietary AI architectures on top of Meta's free infrastructure rather than paying closed-model API tolls.

Mechanically, embedding a 405B-parameter open model into WhatsApp — a platform with billions of users — is a distribution bet of the same kind Google is making with Gemini in Search: reach via existing surfaces rather than standalone apps. For Meta, free Llama-in-WhatsApp also feeds engagement and data within its own ecosystem.

The open-source angle resonates with the week's anti-lock-in theme — the same instinct behind AWS open-sourcing ExtendDB and Mistral's sovereign-compute push. Separately, reports indicate Meta is developing an AI-powered pendant wearable, following OpenAI's wearables move, signaling Meta wants both the software (Llama) and hardware layers. The skeptical read: 'commoditize your complement' is a well-worn strategy, and questions remain about whether Llama's open releases keep pace with frontier closed models like Opus 4.8 and GPT-5.5 on agentic tasks. Watch Llama 4's benchmark positioning and whether WhatsApp's free AI meaningfully shifts user behavior.