GoKawiilLean 4.35.0 adds lake check and four extra kernels including proven-consistent con-leche
The Lean proof assistant's new v4.35.0 release introduces a lake check command that re-verifies developments using leanchecker inside a sandbox, plus a --paranoid mode that runs four additional kernels: nanoda, lean4lean, con-leche, and con-ron. Con-leche, built in Lean itself by Claude models under supervision from Joachim Breitner at the Lean FRO, is formally proven consistent against a set-theoretic model and reportedly runs faster than the official kernel on the full Mathlib library.
GoKawiil's interpretation of the reporting above, not reported fact.
The move follows recent postmortems describing AI agents discovering vulnerabilities in Lean's kernel, suggesting the team is responding to those findings by diversifying and hardening verification paths rather than relying on a single trusted implementation. A kernel proven consistent in Lean, built with AI assistance, could reduce the risk of implementation bugs undermining proof trust, though the developers note remaining risk from the compiler and runtime layers beneath it.
- Lean v4.35.0 adds 'lake check' to recheck proofs with the official kernel in a sandbox
- A new '--paranoid' mode runs four extra kernels, including con-leche, the first Lean-written and proven-consistent kernel
- Con-leche was built with help from Claude AI models and runs faster than the official kernel on Mathlib
Source: leodemoura.github.io — Leonardo De Moura, 2026-10-16
Published there as: “Lean now ships with verified kernels”
Read the original report → The summary and analysis above are GoKawiil's own, written from reporting by the source above. Facts and quotes belong to the original publisher.