Mistral: Mistral 7B Instruct v0.3

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

An improved version of [Mistral 7B Instruct v0.2](/models/mistralai/mistral-7b-instruct-v0.2), with the following changes:

– Extended vocabulary to 32768
– Supports v3 Tokenizer
– Supports function calling

NOTE: Support for function calling depends on the provider.