Axiom Math Releases 43.1M-Line Lean Proof of Finite Simple Group Classification
Axiom Math announced on October 27, 2026 that it is releasing what it calls the first complete computer-checked proof of the Classification of Finite Simple Groups (CFSG). According to the company, swarms of dozens of AI agents wrote approximately 43.1 million lines of Lean code over 24 days.
Axiom Math said on October 27, 2026 that it is releasing what it calls the first complete computer-checked proof of the Classification of Finite Simple Groups (CFSG). The company claims the proof was produced in 24 days by engineers and mathematicians directing swarms of dozens of AI agents, which wrote approximately 43.1 million lines of code in the Lean proof assistant.
"Today, we are releasing the first complete computer-checked proof of the Classification of Finite Simple Groups (CFSG)," the company wrote in its announcement, as reported by Unite.AI.
Key figures
- Output: approximately 43.1 million lines of Lean code
- Time to produce: 24 days
- Workforce: engineers and mathematicians directing swarms of dozens of AI agents
- Proof assistant: Lean
- Announcement date: October 27, 2026
The completeness claim, and the description of the proof as the first of its kind, come from Axiom Math. No model name, benchmark score, pricing or context window figures are attached to the announcement as reported.
What this means
Axiom Math describes the work as a human-directed effort in which AI agents produced the Lean code, with engineers and mathematicians steering the swarms. The scale it reports, approximately 43.1 million lines in 24 days, frames this as a demonstration of agent-driven formal proof engineering rather than a single-model result. Because the proof is written in Lean, it is a computer-checked artifact, which is the basis of the company's claim of a complete formal proof of the CFSG.
Related Articles
OpenAI says it found the Exponential Time Hypothesis is false, alongside two other discoveries
OpenAI announced three new discoveries in mathematics, computer science and physics. The company says one of them shows that the Exponential Time Hypothesis, a longstanding conjecture, is false.
Anthropic recommends consolidating sub-agents into one agent that loads context on demand
Anthropic is advising companies to evaluate whether they can replace fleets of specialized sub-agents with one primary agent that loads context as needed. In one comparison, the company says the single-agent setup ran about twice as fast. A migration kit is available on GitHub.
Anthropic Says Claude Proved Security Properties for 239 Entry Points Across 123 Libraries
Anthropic published research on October 22, 2026, reporting that Claude produced machine-checked formal proofs of security properties at the RISC-V assembly level. The proofs cover 239 fuzz-test entry points across 123 open-source libraries, in a collaboration with Theorem.
OpenAI agents on GPT-Rosalind analyze mouse brain CRISPR atlas in 32 minutes, versus months by hand
OpenAI, PerturbAI and Scripps Research deployed thousands of GPT-Rosalind agents to analyze a whole-brain CRISPR atlas covering millions of mouse cells. The full analysis took 32 minutes; the team estimates manual work would take several months to a year.
Comments
Loading...