A high-performing, industry-standard 7.3B parameter model, with optimizations for speed and context length.

Mistral 7B Instruct has multiple version variants, and this is intended to be the latest version.