The announcement came during the company’s first AI developer conference, marking a strategic move to compete with OpenAI, Google, and emerging players like China’s DeepSeek.
“You can now start using Llama with one line of code,” said Meta Chief Product Officer Chris Cox.
The Llama API, currently in limited preview, will roll out more broadly over the coming months.
While Meta did not disclose pricing, the move reinforces its strategy of offering the Llama models mostly free of charge to promote innovation and deeper integration into its platforms.
Meta also introduced a standalone AI assistant app and confirmed plans to test a paid AI chatbot subscription in Q2, signaling broader commercial ambitions.
Company executives emphasized the flexibility of its offering. “Whatever model you customize is yours to take wherever you want, not locked on our servers,” said Manohar Paluri, VP of AI at Meta.
CEO Mark Zuckerberg positioned Meta’s open approach as a counterweight to industry concentration. “If another model, like DeepSeek, is better at something, you can now combine the best intelligence from multiple models,” he noted.
With the latest Llama release touting lower costs and improved efficiency, Meta is positioning itself as both a cost-effective and developer-friendly alternative in the AI arms race.
The new Llama API reflects Meta’s long-term bet that openness and flexibility will drive adoption, innovation, and engagement across its digital ecosystem.
Would you like a comparison of how Llama’s API stacks up against OpenAI or Google’s APIs?