Meta launches free preview of 'Llama API', ensuring compatibility with OpenAI's API and allowing developers to quickly start developing AI products using Llama
Meta announced the launch of the Llama API , a platform for product development using its in-house AI model Llama, on Tuesday, April 29, 2025. At the time of writing, the platform is in a limited preview stage and can be used for free during the limited preview period.
Llama API | Empowering AI Development with Ease
Everything we announced at our first-ever LlamaCon
https://ai.meta.com/blog/llamacon-llama-news/
In addition to models such as 'Llama 4 Scout' and 'Llama 4 Maverick' released on April 5, 2025, the Llama API also supports 'Llama 3.3 70B' and 'Llama 3.3 8B'. Model fine-tuning and evaluation tools have also been released, and custom models based on Llama 3.3 8B can also be developed.
Users can generate new API keys with one click.
On the usage history screen, you can check token usage by model.
An SDK is already available for development in Python and TypeScript, and compatibility with OpenAI's API is also ensured , allowing developers to quickly develop new products by leveraging the code of existing AI products.
At the time of writing, the Llama API is in limited preview and requires registration on a waitlist to actually use it. Also, during the limited preview, it is available for free.
The Llama API documentation can be found at the following links:
Overview | Llama API
https://llama.developer.meta.com/docs/overview
According to Meta, the total number of downloads of the Llama series has exceeded 1.2 billion as of April 30, 2025. Meta has released a comparative graph with the number of model downloads of rival companies such as OpenAI, Mistral AI, Alibaba, Microsoft, Google, Amazon, and DeepSeek to show its results.
in Software, Posted by log1o_hf