Meta previews an API for its Llama AI models

At his inaugural conference for the Llamacon AI development programmers on Tuesday, Meta announced API for his llam series of AI models: Llam API.

Available in A limited examinationThe API llam enables developers to explore and experiment with products that run different models of LLE, by the meta. Paired with Meta SDKs, developers allows the construction of services, tools and applications aimed at llam. The target did not immediately share the prices of API with Techcrony.

The performance of the API comes as the target looks maintain lead in the fiercely competitive space of the open model. While the lama models were collected to date more than a billion takeoverAccording to the target, rivals like Deepseek and Alibabine Qwen threaten to make the target of effort to establish a far-reaching ecosystem with LLA.

API LLE has tools for fine adjustment and assessment of the performance of the Llam model, starting with Llam 3.3 8B. Customers can generate data, train on it, and then use a meta -ov evaluation package in Llam API to test the quality of their custom model.

Meta said that he would not use information about the customers of Llam API to train their own models of the company, and that the models built using Llam API can be transferred to another host.

For Devs Building at the top of the recently released Llam 4 Meta models, Llama API offers models serving through partnerships with cerebrums and groq. These “early experimental” options “are available on request” to help the developers prototize their AI apps, Meta said.

“By simply selecting a cerebri or groQ name of models in API, you can developers […] Enjoy a simplified experience with all the uses accompanied by one location, “the Meta wrote in a blog post that was provided by Techcrony.”[W]We look forward to expanding partnerships with additional providers to bring even more opportunities to build on the top of LLA. “

The target said he would expand the approach to Llam API “in the coming weeks and months.”