Meta is exploring options for integrating third-party artificial intelligence models into its products, including the Meta AI chatbot.
As reported by The Information (via Engadget), the management of the new Meta Superintelligence Lab division is considering using Google's Gemini model to improve text responses to user queries in search functions.
In addition to Google, Meta is also in talks to use OpenAI models in its apps. A Meta representative said the company is taking a "comprehensive approach" to developing AI products, which includes both partnering with other companies and building its own models.
According to the source, the use of external models is a temporary measure aimed at improving Llama's internal models to remain competitive in the market. It is known that Meta employees already have access to Anthropic's models, which are used for the internal programming assistant.
In parallel, Meta offers high compensation packages to attract AI researchers from Google and OpenAI to its Superintelligence Lab.