Meta Platforms has introduced a new application programming interface (API) to encourage businesses to build AI solutions using its Llama large language models.
Unveiled at Meta’s inaugural AI developer conference, the Llama API positions the company to compete with similar offerings from OpenAI (backed by Microsoft), Google, and emerging low-cost rivals like China’s DeepSeek.
Chief Product Officer Chris Cox highlighted the simplicity of using the API, stating that developers can now deploy Llama models with just a single line of code. APIs are essential for integrating technologies into various products, and they are a major revenue source for companies like OpenAI.
Although Meta didn’t reveal pricing for the API, it announced a limited preview is available to select users, with a broader rollout planned in the coming months. On the same day, Meta also launched a separate AI assistant app and is reportedly preparing to trial a paid AI chatbot service in the second quarter.
Unlike competitors, Meta provides its Llama models largely for free, aiming to stimulate innovation, reduce reliance on rival platforms, and increase user interaction across its social apps. AI VP Manohar Paluri emphasized that developers maintain full ownership of customized models, offering freedom not typically allowed by other providers.
In response to concerns over the high cost of AI development, Meta developers shared cost-cutting methods that boosted the performance of the latest Llama version. CEO Mark Zuckerberg welcomed the rise of competitive players like DeepSeek, noting that their strengths can be leveraged by developers to mix and match AI capabilities across models, fostering a more diverse and powerful AI ecosystem.
