← Back to models
↗↗↗
Mistral
Mistral Large 3
Mistral’s open-weight multimodal flagship with a Mixture-of-Experts architecture.
Overall score
89
open-weightsmultimodalmoe
Context window
256K tokens
Speed
Balanced
Input pricing
$0.50 / 1M input tokens
Output pricing
$1.50 / 1M output tokens
Score breakdown
Capability90
Use-case fit89
Cost efficiency91
Speed83
Reliability86
Agent readiness88
Ecosystem84
Scores combine benchmark signals, product experience, and editorial weighting. Use them as a practical guide, not an absolute truth claim.
Best for
CodingResearch
Works with
Mistral APIself-hosted evaluationdocument workflows
Modalities
textimage
Sources & trust
Officially verified core fields
Official linkSummaryDescriptionModalitiesContext windowPricingPricing pageTool support
Editorial fields such as shortlist guidance, strengths, caveats, and scoring remain clearly separated from official provider data.
Mistral official
Official site · Tier 5 · Apr 9, 2026
Official link
Mistral Large 3 docs
Official docs · Tier 5 · Apr 9, 2026
SummaryDescriptionModalitiesContext windowPricingPricing pageTool support
Mistral Large 3 VerdictLens review
Manual review · Tier 3 · Apr 9, 2026
Best-fit guidanceWorks-with guidanceStrengthsCaveatsOverall scoreScore breakdown
Last verified: Apr 9, 2026
Strengths
- Open-weight positioning makes it strategically interesting for some teams.
- Price-performance looks strong relative to many premium closed models.
Things to watch
- Tooling ecosystem is still smaller than the biggest US vendors.
- Teams should validate downstream hosting and maintenance assumptions.
Best for
Coding copilots & repo execution
Prioritize reliability, diff quality, tool-calling control, and the ability to maintain focus across multi-file edits.
Research synthesis & analyst workflows
Prioritize source grounding, multilingual reading, long-context reasoning, and a retrieval stack that stays inspectable.