Skip to main content

Models / Mistral

Model · Mistral

Mistral

Efficient European models that work well as the model that merges the others.

Mistral builds lean, capable models that get a lot done per dollar. They are a natural fit for the synthesis step in a weave, where the job is to read several answers and combine them well.

The lineup

ModelBest forPlan
Mistral Medium 3.1The default free judge that merges model outputs.Free
Mistral Large 2512A stronger Mistral model on the paid plans.Paid

What it is good for

  • Synthesizing and judging several model answers
  • Efficient, cost-effective general work
  • A European-built perspective in the mix

In a weave

Mistral Medium 3.1 is the default judge in many weaves: the model that reads everyone’s answer and produces the merged result. It is free, which keeps the synthesis step cheap.

Mistral Medium 3.1 is included free and serves as the default judge. Mistral Large is on the paid plans.

Common questions

What does the judge model do?

In a weave, the judge reads the answers from the responding models and merges or ranks them into one result. Mistral Medium 3.1 is the free default for that role.

More models

Try Mistral in a weave

Run it next to other models on your own task. Free to start, no card.

Get started