META releases a new tool that, as he hopes, will encourage developers to use their family of LLAMA models for his next project. In his inauguration Llamacon event On Tuesday in Menlo Park, the company announced the API Llama. Available in the form of limited free preliminary viewing, starting from today, the tool gives developers a place for experiments with Meta's Models, including recently released Llama 4 Scout and Maverick SystemsThe field also allows you to easily create new API keys that developers can use for authentication purposes.
“We want it to be easier for you to quickly start building with Llama, as well as provide you with full control over your models and scales, not being locked on the API,” the blog published in the blog published during the event said. To this end, the initial release of the Llama API includes tools that DEVS can use to accurately configure and evaluate their applications.
In addition, Meta notes. He will not use user tips and answers of models to teach his own models. “When you are ready, the models that you build on the API Llama, you need to take with you wherever you want to place them, and we do not keep them blocked on our servers,” the company said. META expects to expand the tool for more users in the coming weeks and months.
Despite the fact that META's LLAMA models were loaded more than one billion times, the company, as a rule, is not considered as a leader in the space of artificial intelligence just like Openai and Anthropic. This does not help to confront the perception that the company was caught Gaming Lmarena So that Llama 4 models look better than they really were.