Context131k
This is the 400B class of Llama3, with a 128k context window. This 405B instruct-tuned version is optimized for high-quality dialogue and demonstrates strong performance compared to leading closed-source models, including GPT-4o and Claude 3.5 Sonnet.