Meta has developed and publicly released the Meta Llama 3 family of large language models (LLMs), a collection of pre-trained and fine-tuned generative text models ranging in scale from 8 billion to 70 billion parameters. The fine-tuned versions, called Llama 3, are optimized for dialogue use cases. Llama 3 models perform well on the benchmarks we tested, and in our human evaluations for helpfulness and safety, are on par with popular closed-source models.

This offer enables access to Llama-3-8B inference APIs and hosted fine-tuning in Azure AI Studio. Azure AI Studio is the perfect platform for building Generative AI apps. AI Studio comes with features like playground to explore models and Prompt Flow to for prompt engineering and RAG (Retrieval Augmented Generation) to integrate your data in your apps.

The inference APIs are billed on a pay-as-you-go basis for the input and output tokens when scoring with the model. The fine-tuning jobs for models that support fine-tuning are charged by the hour and inference for fine-tuned models is also billed by the tokens used plus an hourly hosting fee.

This offer is integrated with Azure AI Studio. You will be asked to navigate to Azure AI Studio to subscribe to this offer and use the models.