Jiang and colleagues: Mixtral 8x7B sparse mixture-of-experts model

What changed

Released and evaluated Mixtral 8x7B, routing each token through two of eight feed-forward experts per layer with a reduced active-parameter count.

Scientific scope

Released and evaluated Mixtral 8x7B, routing each token through two of eight feed-forward experts per layer with a reduced active-parameter count. Scope limitation: The company announced the model in December 2023; this record uses the January 2024 paper year. Reported benchmark comparisons are setup-specific.

Evidence

The retained primary-source abstract, bibliographic record and stated review scope support this provisional assessment. No independent experimental replication by WRS is claimed. The company announced the model in December 2023; this record uses the January 2024 paper year. Reported benchmark comparisons are setup-specific.

Contributors

• The named authors listed in the linked research record · SCIENCE: Complete source byline retained in the research record; human research direction, method development and evaluation. Individual labor shares are not measured.

• Jiang and colleagues: Mixtral 8x7B sparse mixture-of-experts model — original paper
• Jiang and colleagues: Mixtral 8x7B sparse mixture-of-experts model — official company source

Provisional WRS assessment, edition 2026-10-16.1. Independent expert calibration is pending. This grade does not certify correctness or assign ownership.

Full scientific record · Browse contribution grades · Request a correction