Mixtral 8x7B Instruct
Open weightsOpen weightsText56B total / 14B active (MoE)
Mistral core family model by Mistral.
About the Mistral core family
Mistral's main line: open Small models, the Mixtral MoE series that popularized sparse experts, Pixtral vision models, and commercial Medium/Large tiers up to Mistral Large 3 and Medium 3.5.
All Mistral core modelsHow this one is configured
This configuration mixture-of-experts with ~56B total / ~14B active parameters per token.
Who makes it
MistralFranceParis-based lab known for efficient open-weight and commercial models.More from Mistral core
- Mistral 7B Instruct
- Mistral Large (Feb '24)
- Mistral Large 2 (Jul '24)
- Mistral Large 2 (Nov '24)
- Mistral Large 3
- Mistral Medium
- Mistral Medium 3
- Mistral Medium 3.1
- Mistral Medium 3.5
- Mistral Saba
- Mistral Small (Feb '24)
- Mistral Small (Sep '24)
- Mistral Small 3
- Mistral Small 3.1
- Mistral Small 3.2
- Mistral Small 4 (Non-reasoning)
- Mistral Small 4 (Reasoning)
- Mixtral 8x22B Instruct
- Pixtral Large