Llama 3

Llama 3 is Meta's open-source large language model family, available in 8B and 70B parameter sizes, representing a significant leap in open-weight model capabilities. Trained on over 15 trillion tokens of publicly available data, Llama 3 demonstrates strong performance in reasoning, coding, and creative tasks, competing with proprietary models across major benchmarks. Released under a permissive license that allows commercial use, Llama 3 has become the foundation for thousands of community f...

Threads