The authors developed Bolzano, an open-source multi-agent system for automated mathematical proof search.
Multi-agent LLM system solves hundreds of open math problems without human steering
Bolzano, an open-source system combining parallel prover agents with a verifier, answered four questions from STOC 2026 papers and solved about 200 open problems across combinatorics and theoretical computer science.
Academic
Adrián Zámečník · Matěj Kripner · Martin Koutecký · Martin Balko · Jan Grebík · Pavel Hubáček · +2 more
Charles University · Czech Academy of Sciences
Research Digest··3 min read
The authors built Bolzano, a multi-agent system where parallel prover agents propose informal proofs and a verifier agent checks them, all maintaining a human-readable research state.
Why this paper
From Charles University and Czech Academy of Sciences
In one line
Bolzano, a multi-agent system, solved about 200 open problems from 3800 attempts without problem-specific guidance, including four from STOC 2026 papers.
What we could check
- ·No code link found
- ·No weights link found
- ·No dataset link found
- ·No compute details found
- ✓Limitations stated by the authors (3 noted)
- ·No benchmark numbers found
Observed from the paper text and links we have. Absence here means we did not find it, not that it does not exist.
§