Mixtral 8x7B

Model ● Covered in 4 stories + Follow

Mixtral 8x7B is a sparse mixture-of-experts language model released by Mistral AI under an open Apache 2.0 license, featuring 46.7 billion total parameters but using only 12.9 billion per token during inference. The model matches or outperforms GPT-3.5 and Llama 2 70B on most benchmarks while running approximately 6 times faster, and is accessible through Mistral's API platform and integrated across Hugging Face's ecosystem including Transformers and Inference Endpoints.

Updated 8 August 2026

Specifications

No specifications recorded yet.

Latest developments

Au Large Mistral AI · 2 years ago · 53

Mixtral of experts Mistral AI · 2 years ago · 42

Mixture of Experts Explained Hugging Face · 2 years ago · 39

Welcome Mixtral - a SOTA Mixture of Experts on Hugging Face Hugging Face · 2 years ago · 5

Timeline

Month Quarter Year

February 2024

26 Feb Au Large

December 2023

Mistral AI releases Mixtral 8x7B, an open-weights mixture-of-experts language model Model release

11 Dec Mixtral of experts
11 Dec Mixture of Experts Explained
11 Dec Welcome Mixtral - a SOTA Mixture of Experts on Hugging Face

Relationships

Products & technology

Integrated with Hugging Face · 1 source
Mistral AI develops this model · 1 source
Mistral develops this model · 1 source

Every AI story that matters,
in your inbox by 8am.

TLDRocket reads all relevant sources, removes duplicate coverage, and summarises the day in two minutes. Follow companies and topics for alerts, or get the briefing in Slack. Free, no spam, unsubscribe anytime.