Mistral's multimodal model. Native image understanding meets strong EU-rooted open weights.
Pixtral has become a mainstay of modern AI workflows. Whether you're a solo creator or part of a larger team, it earns its place by doing one job exceptionally well — and playing nicely with the rest of your stack.
Who it's for: Developers who want open, self-hostable multimodal models with strong document understanding.
Reason over images and documents natively.
12B and Large sizes released openly.
Charts, screenshots, and PDFs.
Agentic workflows with vision context.
Pixtral is the open multimodal model to watch when you need vision without vendor lock-in.