Meta Llama 3.1 has 405B parameters and supports a context length of 128k, multilingual capabilities, and excels in tasks like code generation and complex reasoning. It was trained on 15 trillion tokens using over 16,000 H100 GPUs. This release includes tools for easy integration and is designed to be competitive with leading models like GPT-4 and Claude 3.5. It also introduces features like Meta AI Imagine and the Llama Stack API
Llama 3.1 405B was released on July 23, 2024.
As a multilingual AI mode, Llama 3.1 405B supports English, German, French, Italian, Portuguese, Hindi, Spanish, and Thai.
The Llama 3.1 405B context length is 128k characters.
Yes as an open source model, Llama 3.1 405B can be fine tuned and deployed.
LLama 3.1 405B is available at Meta partners including AWS, Azure, Google Cloud, Dell, Groq, Databricks and most major computing providers.