Mistral Large 4
Language modelCard summarizedState-of-the-art, open-weight, general-purpose multimodal model
Public preview● Re-checked 12 Oct 2026 · 34 facts · 0 corrections · 2 unconfirmedSource: Introducing Mistral Large 4
Use it via API
$722 /month
$14.44 per 1K requests
RAG over docs: 12K in · 600 out · 50K requests/mo · 30% cached · change
Where to call it
- Mistral API (Mistral Studio)
mistral-large-4
Good for
- Agentic coding and agentic business workflows
- Cybersecurity: vulnerability research, incident response, malware analysis, detection rules
- Multimodal understanding and visual grounding
Overview
Mistral AI · 07Large, front
Large 4
State-of-the-art, open-weight, general-purpose multimodal model
- Cost
- $1.36 in
- Context
- 1M
- Input / output
- $1.36 / $4.18
- Max output
- —
- Knows up to
- —
- Released
- 6 Oct 2026
Mistral's largest model, released as a public preview. A natively multimodal mixture-of-experts model with about 1 trillion total and 52B active parameters, combining instruct and reasoning in one model. Weights are due by the end of October 2026. launch post
mistral-large-4- Input / output per 1M
- $1.36 / $4.18
- Context window
- 1M
- Max output
- —
- Knows up to
- —
- Reasoning
- Hybrid instruct and reasoning
- Weights
- Closed
About this card. A model card is the maker's own account of a model: what it's for, how it was trained and tested, and where it falls short. This page condenses Mistral AI's Introducing Mistral Large 4. Every claim carries its source page.