Mistral 7B is a 7.3 billion parameter language model that represents a significant advancement in large language model capabilities.

• English language tasks
• coding tasks

Mistral 7B outperforms the 13 billion parameter Llama 2 model on all tasks and surpasses the 34 billion parameter Llama 1 on many benchmarks.

• One Billion Tokens with Revolutionary LONGNET Architecture