Meta visualizes API for its Llama AI models

At its opening conference on Llamacon AI developer on Tuesday, Meta announced API about its AI series of Llama: The Llama API.

Available in Restricted examinationThe Llama API allows developers to explore and experiment with products powered by different Llama models for meta. Combined with Meta SDK, it allows developers to build Llama -oriented services, tools and applications. Meta did not immediately share API prices with TechCrunch.

The API invention is coming as Meta seems to maintain a lead in a fiercely competitive open space for an open model. While Llama’s models have gathered So far more than a billion downloadsAccording to Meta, rivals like Deepseek and Qwen of Alibaba threaten to raise Meta’s efforts to establish a distant ecosystem with llama.

Llama’s API offers fine -tuning and evaluation tools of Llama models, starting with Llama 3.3 8B. Customers can generate data, train on it and then use the Meta Meta API assessment package to test the quality of their personalized model.

Meta said it would not use Llama API client data to train the company’s own models and that models built using Llama API can be transferred to another host.

For the construction of Devs at the top of the recently issued Meta Llama 4 models, Llama’s API offers options for servicing models through partnerships with CEREBRAS and GROQ. These “early experimental” options are “available on request” to help developers prototyping their AI applications, Meta said.

“By simply choosing the names of models or models of GROQ in API, developers can … enjoy a rationalized experience with all traced use in one place,” Meta wrote in a blog post provided to TechCrunch. “(W) We look forward to expanding partnerships with additional suppliers to bring even more opportunities to build Llama.”

Meta said it would extend access to Llama’s API “In the coming weeks and months.”