Project Timeline
Key milestones in the development of IMProofBench
End of March 2025
Whitepaper Publication
First version of whitepaper released with basic benchmark design and theoretical framework.
Completed
May 2025
Community Engagement
Exchange of ideas with other benchmark organizers, refining benchmark design based on community feedback and existing evaluation frameworks.
Completed
June 4, 2025
First Brainstorming Session
Question brainstorming session at the 2025 Helvetic Algebraic Geometry Seminar, gathering initial problem contributions from the community.
Completed
June 13, 2025
Website Launch
improofbench.math.ethz.ch goes online for the first time, providing the submission platform for mathematical researchers.
Completed
June 19, 2025
MFO Workshop Session
Question brainstorming session at the MFO Workshop "Recent Trends in Algebraic Geometry", expanding problem collection.
Completed
July 29, 2025
Review System Added
System for problem reviews integrated into the website, enabling peer review process for submitted questions.
Completed
August 29, 2025
Review Invitations & Grading System
Review invitation system and grading interfaces go online, enabling comprehensive peer review and AI evaluation grading workflows.
Completed
Early September 2025
Pilot Evaluation Progress
Over 20 questions accepted and 8 questions fully graded. Continuing to expand the benchmark with community contributions and evaluations.
Completed
Mid-September 2025
Paper Write-up & Analysis
Complete pilot evaluation of 25+ questions and begin comprehensive analysis of results and AI performance patterns for paper write-up.
Completed
September 24, 2025
ICLR Submission
ICLR 2026 paper submission deadline - submit proof-of-concept paper with pilot results and benchmark methodology.
Completed
September 30, 2025
Preprint Publication
First results paper published on arXiv, authored by the 5 core team members and 28 contributors, presenting pilot evaluation results and benchmark methodology.
Completed
Timeline Legend
Completed
Past milestones
Upcoming
Future goals
Major milestones