SAN FRANCISCO: In an effort to entice companies to develop AI products more effortlessly using its Llama artificial-intelligence models, Meta Platforms unveiled an application programming interface.
Llama API, announced in the company‘s first-ever AI developer conference, will enable Meta to compete against APIs provided by other model creators such as Microsoft-funded OpenAI, Alphabet‘s Google and new low-cost challengers like China‘s DeepSeek.
“You can now begin to use Llama with a single line of code,” chief product officer Chris Cox declared in a keynote presentation onstage.
APIs enable software developers to tailor and integrate a bit of technology into their own products very quickly. APIs are the company’s main source of revenue for OpenAI.
Meta, which rolled out the newest version of Llama earlier this month, did not disclose any pricing information for the API.
In a statement, it announced the new API was out as a limited preview for select customers and would be rolled out broadly in weeks to months.
The company also launched a standalone AI assistant app separately on Tuesday. It plans to test a paid subscription service of its AI chatbot in the second quarter, Reuters reported in February.
Meta makes its Llama models widely available for free to developers to use, a move CEO Mark Zuckerberg earlier said will pay dividends in the form of new products, reduced reliance on potential competitors and increased activity on the company‘s main social networks.
“You have complete control over these bespoke models, you own them in a way that you can’t with other propositions,” Manohar Paluri, an AI vice president, said at the event. “Whatever model you tailor is yours to use wherever you wish, not stuck on our servers.”
DeepSeek, also releasing partially open-source AI models, ignited a January stock selloff as investors feared the expense of AI model creation required by leading U.S. companies.
During the conference, Meta developers also discussed new methods they employed in order to effectively lower costs and enhance the efficacy of its newest Llama generation. Zuckerberg was welcoming more competition that would shift the competitive space away from monopolization by the few leaders.
“If another model, such as DeepSeek, does something better, then now as developers you can take the best of the intelligence of the various models and create precisely what you require, which I believe is going to be really powerful,” Zuckerberg stated.