Hugging Face & Mistral Release Mixtral 8x22B

A Smarter, Lighter Model

Mixtral 8x22B is a sparse MoE (mixture-of-experts) model, meaning only a fraction of its parameters are used per query — offering top-tier performance with reduced costs.

Highlights:

• Open-source with Apache 2.0
• Efficient inference with 2 active experts
• Competitive with GPT-4 on many tasks

Explore:

• Hugging Face Model Page