Mistral AI
Mixtral 8x7B Instruct
LegacyMixtral 8x7B Instruct is a sparse mixture-of-experts open model noted for strong quality per active parameter and efficient inference vs dense models of similar capability.
Open weights MoE LLM · Release — · Apache-2.0
Newer version: Mixtral 8x22B
Updated 1 day ago · Verified Apr 2026 · Score 78
Decision summary
Why teams reach for it, where it fits, and what to watch for — before you dive into specs.
Why teams choose it
- Throughput benefits depend on expert routing implementation—benchmark your serving stack.
- Great mid-tier alternative when Mistral Large is overkill.
Best use cases
- Use this when self-hosted chat for EU data residency
- Use this when cost-aware coding assistants
Tradeoffs
- MoE serving is trickier than dense models—ops matter.
- Not the latest frontier—compare to newer Mixtral/Mistral SKUs.
Technical details
Modalities, benchmarks, and release context.
Modalities
What goes in and what comes out.
- Inputs
- text
- Outputs
- text
- Capabilities
- MoE, multilingual, coding
Benchmarks snapshot
Structured JSON for reproducible comparisons.
No benchmark data yet — see comparisons for relative performance.
Family lineup
Explore other versions in this family after you have the headline on this model.
Continue exploring
A short set of comparisons, nearby models, and links to go deeper — without repeating the same paths.
Compare with
Learn & build
Tools and curated destinations (max four).