Metraipolis › archive › dc_efe3d5e118ea

Stellar Colosseum: A Many-Agent Harness for Long-Horizon Research in Mathematics and Theoretical Computer Science — arXiv:2609.15983

Source: arXiv (en)
Original: https://arxiv.org/abs/2609.15983
License (as archived): CC0-1.0
Archived: 2026-10-15T22:48:10.411Z
SHA-256 of the archived text: 12fb8f1522335529a8ac48430a1c23c8ada6f87ddd38df272b46803d0810c32f
Stable URL of this page: https://metraipolis.com/documento/dc_efe3d5e118ea

How to cite

Stellar Colosseum: A Many-Agent Harness for Long-Horizon Research in Mathematics and Theoretical Computer Science — arXiv:2609.15983 — arXiv (en), https://arxiv.org/abs/2609.15983 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_efe3d5e118ea (Metraipolis archive dc_efe3d5e118ea, sha256 12fb8f1522335529a8ac48430a1c23c8ada6f87ddd38df272b46803d0810c32f)

BibTeX

@misc{dcefe3d5e118ea,
  title = {Stellar Colosseum: A Many-Agent Harness for Long-Horizon Research in Mathematics and Theoretical Computer Science — arXiv:2609.15983},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.15983},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_efe3d5e118ea; sha256 12fb8f1522335529a8ac48430a1c23c8ada6f87ddd38df272b46803d0810c32f},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.15983. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Language models can produce plausible short proofs, but may still be unreliable on long-horizon research problems, where progress depends on a sequence of uncertain and interdependent decisions. We introduce Stellar Colosseum, a model-agnostic harness for allocating inference across research in mathematics and theoretical computer science. Colosseum explores alternative strategies before proof construction, uses a readiness gate to decide when a route is mature enough to decompose, represents the proof plan as interdependent section-level subproblems, and routes verifier findings back to the affected part of the argument. Across these stages, it generates candidates in parallel, attacks them with targeted falsification, and combines candidates and their critiques into a single research artifact through overlapping random-sample tree aggregation. The Colosseum workflow has been integrated into Google Antigravity's Teamwork framework as the Long Proof pattern. We demonstrate the capabilities of Colosseum through open-ended research and evaluations on theorem-proving and competitive programming benchmarks. Using Colosseum with Gemini 3.1 Pro, we obtain several new results that address open problems arising from papers published at top venues such as FOCS and JMLR. On TCS-Bench, a benchmark of research-level theorem-proving tasks drawn from papers published at FOCS, STOC, and SODA, Colosseum achieves 71.0% accuracy using Gemini 3.1 Pro and Gemini 3.7 Flash. In a separate Codeforces evaluation using Gemini 3.1 Pro, the proof-oriented pipeline with execution feedback solves 218 of 222 problems.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-15 in cs.AI).
Categorie arXiv: Artificial Intelligence (cs.AI) ; Computation and Language (cs.CL); Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2609.15983 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-15T22:29:03.380Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-15?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).