News

On January 29, 2024, Meta released the “Code Llama 70B” model, which is based on the 70 billion parameter “Llama 2 70B” model and additionally trained with code-specific datasets.