# Researchers outline a multi-session workflow for Olympiad agents

> HOA combines independent review, structured knowledge notes and formal verification.

- Canonical URL: https://newsverum.com/en/news/researchers-coordinate-multi-session-olympiad-agents
- Publisher: Newsverum
- Section: Technology
- Country: China
- Published: 2026-10-13T00:51:41+00:00
- Updated: 2026-10-13T00:51:41+00:00
- Event / data period: 2026-10-09
- Other editions: [Türkçe](https://newsverum.com/tr/news/researchers-coordinate-multi-session-olympiad-agents.md), [العربية](https://newsverum.com/ar/news/researchers-coordinate-multi-session-olympiad-agents.md), [Français](https://newsverum.com/fr/news/researchers-coordinate-multi-session-olympiad-agents.md), [Español](https://newsverum.com/es/news/researchers-coordinate-multi-session-olympiad-agents.md), [Deutsch](https://newsverum.com/de/news/researchers-coordinate-multi-session-olympiad-agents.md), [Português](https://newsverum.com/pt/news/researchers-coordinate-multi-session-olympiad-agents.md)

Researchers introduced HOA, an open workflow that coordinates multiple agent sessions on difficult mathematics and science problems. It sends candidate solutions to an independent reviewer, supplies missing domain knowledge through a structured wiki, and checks Lean-formalized proofs with a trusted verifier.

The authors report solving 672 PutnamBench problems and 251 problems on Lean-Eval with HOA. Experiments also cover the official 2026 mathematics, physics, chemistry and programming Olympiad problem sets; scoring varies by field and uses human review, code tests or an LLM judge. The findings are limited to these tasks and evaluation procedures and do not establish general scientific-discovery ability.

## Data

### Measure

PutnamBench result in the paper's evaluation

Methodology: Values come from the paper's stated benchmark experiments; they are not field measurements or general performance guarantees.

| Measure | Reported value | Experimental scope |
|---|---|---|
| PutnamBench problems solved with HOA | 672 | PutnamBench result in the paper's evaluation |

Sources: <https://arxiv.org/html/2610.13632v1> · Download data: [JSON](https://newsverum.com/api/v1/articles/researchers-coordinate-multi-session-olympiad-agents/data/reported-result-1?lang=en&format=json) · [CSV](https://newsverum.com/api/v1/articles/researchers-coordinate-multi-session-olympiad-agents/data/reported-result-1?lang=en&format=csv)

### Measure

Lean-Eval result in the paper's evaluation

Methodology: Values come from the paper's stated benchmark experiments; they are not field measurements or general performance guarantees.

| Measure | Reported value | Experimental scope |
|---|---|---|
| Lean-Eval problems solved with HOA | 251 | Lean-Eval result in the paper's evaluation |

Sources: <https://arxiv.org/html/2610.13632v1> · Download data: [JSON](https://newsverum.com/api/v1/articles/researchers-coordinate-multi-session-olympiad-agents/data/reported-result-2?lang=en&format=json) · [CSV](https://newsverum.com/api/v1/articles/researchers-coordinate-multi-session-olympiad-agents/data/reported-result-2?lang=en&format=csv)

## Sources

- [arXiv preprint by Tsinghua University, NVIDIA and collaborating researchers](<https://arxiv.org/html/2610.13632v1>) (Source checked: 2026-10-13; Published: 2026-10-09)

---

Cite the canonical URL above. [Editorial standards and corrections](https://newsverum.com/en/editorial)