Anthropic has Claude generate formal verification proofs automatically
What happened
On 23 October 2026, The Decoder reported that Anthropic's Frontier Red Team worked with Theorem to have Claude turn crash tests for 123 open-source libraries in OSS-Fuzz into mathematical problems and analyse the compiled machine code, generating about 180 million lines of proof code within weeks, independently checked by Rocq. The report says the goal is for Claude to generate formal safety proofs for existing code automatically.
Written by AI from the reports · updated 3 h ago
Timeline
Follow the reports to see the event from each side.
- The DecoderSelectedAnthropic's Claude now writes code and proves it safe automatically
Anthropic had Claude generate mathematical proofs that existing code is correct, producing roughly 180 million lines of proof code for 123 open-source libraries. Along the way, the model caught bugs that years of testing had missed. Anthropic's goal: making formal verification with AI as routine as standard testing.
Heat of this event
Not enough continuous observations yet to draw a trend.