AlphaProof Code at a glance
Google DeepMind introduced AlphaProof Code on November 10, 2026: an agentic framework whose agents generated over six million lines of Lean to prove key security properties of Signal's actual protocol code in four weeks, a first at this depth for a widely deployed protocol. The artifacts are open-sourced.
What changed with AlphaProof Code
Google DeepMind announced AlphaProof Code on November 10, 2026, an agentic framework for formal verification, and used it on one of software security's long-standing grand challenges: proving that the code behind the Signal protocol, which protects private conversations in Signal and Google Messages, actually delivers its specified security properties. DeepMind says this is the first time AI has formally verified a widely deployed cryptographic protocol at this depth, and the first time such an implementation has been formally verified by any method.
The framework splits the work like a research team: human experts collaborate with a mix of models to turn Signal's security goals into formal specifications and review every one, then agents powered by Gemini Carbon write the proofs, generating over six million lines of Lean, the language whose checker mathematically validates each step. The agents also audited their own proof and flagged the areas needing human review, about one percent of the codebase. The whole verification took four weeks, work DeepMind says historically takes expert teams years, and as a control the team reintroduced two real, previously fixed bugs into an isolated copy of Signal's code: in both cases the system produced a machine-checked proof that the buggy version was insecure.
The proofs cover Signal's core cryptographic code down to the compiled code running on devices, including protections for messages before and after a device is temporarily compromised, rather than stopping at the protocol's design, where most earlier verification efforts stopped. DeepMind has open-sourced the formal verification artifacts for community use and published the technical paper alongside the announcement. AlphaProof Code is a research framework, not a product: DeepMind announced no API, pricing or general availability for it.
Where to compare AlphaProof Code
Independent leaderboards and official resources for comparing this model yourself. Rankings change often; each site shows its own current data.
Why AlphaProof Code matters for businesses
Why it matters: formal verification is the strongest assurance software can carry, and it has been too slow and expensive for almost everything that matters. If AI agents can do it in weeks, the economics change for the critical systems agencies and their clients depend on, from messaging to medical and industrial software, and 'the AI checked it' starts to mean something provable instead of a promise.
Get AI News by email
Verified AI stories, each with its official source, on Monday and Thursday mornings. Add breaking news to hear about the biggest ones as soon as we publish them.
Every story links its official source
Free, in English or Spanish
Unsubscribe with one click, any time
Sources for this story
- Formal Verification of Signal's Protocol with Gemini (opens in a new tab)Official
Google DeepMind · deepmind.google
- AlphaProof Code: verifying Signal's security at the machine-code level (technical paper and artifacts) (opens in a new tab)Official
Google DeepMind · google-deepmind.github.io
Last verified: . IA Compa summarizes public announcements for business readers. We are not affiliated with the companies mentioned. Always confirm details, pricing, and availability with the official source.
- #alphaproof-code
- #formal-verification
- #signal-protocol
- #ai-agents
- #software-security
- #lean
- #cryptography
