Leanstral: Open-Source foundation for trustworthy vibe-coding
Source
mistral.ai
Date
Context
Leanstral targets coding tasks where a proof checker can verify whether a proposed result satisfies a formal statement. Mistral AI designed the model for Lean 4, a language and proof assistant used to express mathematical claims and software properties. Its release account describes generating candidate proofs and using Lean to check them, with evalA repeatable test for AI quality — a set of tasks plus scoring — used the way software teams use test suites, because model output is too variable to judge by eyeballing.Full definition → based on changes in a formal mathematics repository. This offers a route to automating proof engineering where requirements have been formalized; the evidence concerns that specialized workflow.
Terms in this piece · Glossary
eval — A repeatable test for AI quality — a set of tasks plus scoring — used the way software teams use test suites, because model output is too variable to judge by eyeballing.
AI agent — An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.