Meta explores using competitors’ AI models to boost Meta AI
Recent reports indicate Meta is considering integrating competitors’ models — including Google Gemini, OpenAI models and Anthropic — to power or supplement Meta AI and features across its apps. Meta says it is taking an “all-of-the-above” approach: partnering with other companies while continuing to build its own Llama models. The use of external models appears intended as a temporary measure to accelerate improvements to Meta’s in-house models and stay competitive.
Key points
- Meta’s Superintelligence Lab has reportedly discussed adding Google Gemini into Meta AI to improve conversational search responses.
- Meta has also reportedly explored using OpenAI models and already gives employees access to Anthropic’s models for internal tools like a coding assistant.
- Meta frames this as a short-term strategy to improve Llama and hire top AI talent to build longer-term capabilities.
Why this matters
Relying on rival models could speed feature improvements and help Meta compete with Google and OpenAI on AI performance. However, it raises questions about product consistency, data sharing, and long-term dependence on external providers.
Sources
- Business Insider: Meta and Anthropic models used in internal tools
- AINvest: Meta’s 2025 AI roadmap and competitive positioning
Would you prefer companies to build all AI in-house or to mix & match best-in-class models? Share your thoughts below.