Quanta AI Labs

Google DeepMind · Agentic formal verification

AlphaProof Code

AlphaProof Code is a multi-LLM agentic framework for formal verification, in which human experts specify security goals and AI agents write Lean proofs. Agents powered by Gemini Carbon generated over six million lines of Lean to formally verify the compiled code of Signal's cryptographic protocol, with humans reviewing only about one percent of the codebase, and the work took four weeks. The team reintroduced two previously fixed bugs and the system produced machine-checked proofs that the buggy versions were insecure. DeepMind says it is open-sourcing its formal verification artifacts, and the technical paper is linked.

Open the source ↗

Similar Agents & Automation releases

  • Claude Meetings

    Anthropic · Meeting Agent · Week 47, 2026

  • PCBRouteAgent

    Yue Zhang, Jilei Sun, Liqiang Jing, Yijie Wei, Ruohan Yang · PCB Routing Agent · Week 47, 2026

  • BookLeme (读了么)

    bookleme (GitHub) · Book-to-Skill Compiler · Week 47, 2026

  • pwner

    0x45dd (GitHub) · Pentest Agent · Week 47, 2026

  • OpenRoboKit

    OpenRoboKit (authors not stated) · Robotics toolkit · Week 46, 2026

  • minecraft-agent-harness

    yc-software · Minecraft agent harness · Week 46, 2026

More from Google DeepMind

Published by Quanta AI Labs, an AI agency in India. Browse every week in AI News Weekly.