Anthropic and Theorem use Claude to automate code safety proofs

A significant shift in software engineering has arrived as artificial intelligence begins to handle the rigorous task of proving code correctness.

Historically, formally verifying that software runs without errors was an incredibly expensive process. A team of specialists required twenty person-years just to verify a small operating system kernel. Consequently, this method was reserved for high-stakes applications such as avionics. Recently, Anthropic’s Frontier Red Team collaborated with Theorem to change this dynamic. They tasked their model, Claude, with taking over the proof generation work. In a matter of weeks, the model produced roughly 180 million lines of mathematical proof code. This effort covered test programs from 123 open-source libraries within Google’s OSS-Fuzz security project.

The approach differs from standard AI coding tools. Instead of writing new software from scratch, Claude verified existing code. It translated crash tests into mathematical problems and analyzed compiled machine code to produce extensive proofs. An independent verification program named Rocq automatically checks these proofs. Anthropic states that the process requires minimal human review. The researchers describe the system as fully autonomous, although the underlying formal foundations were built by human engineers at Theorem.

The model also demonstrated a strong ability to identify hidden flaws. While generating proofs, Claude encountered hundreds of bugs and was permitted to fix them. Many of these issues would have passed conventional testing because they only appear under specific conditions, such as unusually large input files. During the verification of a widely used PDF reader, the team found as many security bugs in one week as are typically reported for that software in an entire year. In another instance, the AI caught a logical error that had remained undetected for more than a decade.

However, the team notes that formal verification does not guarantee software is entirely error-free. The proofs hold only under strict conditions, including a specific processor architecture, single-threaded programs, and inputs under 64 MB. In practice, these libraries often run on different hardware and configurations. The compute cost is also high, with individual proofs running up to hundreds of millions of tokens. Anthropic did not report the total project costs, but a smaller test run showed that proofs are reusable and newer models are becoming more efficient. Opus 5.5 solves half the proofs with about 40 percent fewer tokens than its predecessor.

This project strengthens the argument for software automation. If AI can write code and guarantee its correctness under defined conditions, the focus shifts from whether AI can program to whether it can prove safety. Anthropic expects this form of verification to become as routine as standard testing. Theorem’s roadmap already targets critical infrastructure components like glibc, OpenSSL, and cURL.

For the Pacific Northwest tech sector, this development suggests a future where AI-generated software is more trusted and requires less manual review. This could reduce costs across the entire coding process and accelerate the deployment of reliable software.