Llama 2 is a family of open-weight transformer-based language models developed by Meta and released in July 2023, available in 7B, 13B, and 70B parameter variants. The models feature a 4096-token context window and were trained on 2 trillion tokens with a September 2022 data cutoff. Both pretrained foundation models and instruction-tuned "Chat" versions are available, with the latter optimized through supervised fine-tuning and reinforcement learning from human feedback for improved safety and dialogue capabilities.