AI

OpenFrontiers: A Collection of Open Research Problems and a Benchmark for Scientific Rediscovery

Read full article at source ↗

Based on a feed excerpt; full article content has not been acquired. Source evidence retrieved 30/10/2026, 06:48:24 UTC.

Briefings

OpenFrontiers is a cross-disciplinary resource combining 12,151 open research problems with a benchmark of 200 previously solved problems. It evaluates seven LLM agents on the HQ-50 subset using formal proofs under closed-book, frozen-retrieval, and live-web conditions. The study found that closed-book rediscovery is rare, occurring in at most 7 of 40 resolved problems, while live-web success primarily reflects retrieval of published solutions.

Key points

• OpenFrontiers contains 12,151 open research problems with problem statements, context, progress, and other information.
• The benchmark includes 200 resolved problems, 90 still open, each with a Lean~4 statement checked by automated probes and two independent reviews.

Context

This resource addresses the lack of a comprehensive collection of open research problems and the difficulty in evaluating them on research-level challenges. It provides a foundation for measuring research-level rediscovery and studying progress toward solving currently open scientific problems.

Limitations

Evaluation results are limited to the HQ-50 subset of 200 resolved problems. Closed-book rediscovery is rare, occurring in at most 7 of 40 resolved problems. Live-web success mostly reflects retrieval of published solutions rather than independent rediscovery.

Generated by Bittrees-hosted qwen35-2b sha256:aaf42c8b7c3c on 30/10/2026, 09:56:29 UTC. Based on publisher-provided feed evidence; may not cover the full original article. Check the source for details.

View archived briefing on IPFS
bafybeigia73lfwodjkbq6hdkez6qdn2qhp7ieyt25m2wd4gru2b55265qu