- axis
- math
- verification
- det-gold
- whatItTests
- A benchmark for measuring mathematical progress from today's models toward artificial superintelligence through fourteen parameterised construction families drawn from open mathematical problems; each submitted object is checked automatically for validity and assigned an uncapped relative quality score against a published frontier or construction baseline, with compact certificates keeping large constructions verifiable.
- saturation
- open
- patternLabFit
- An unsaturated math benchmark with an uncapped quality score, rather than a binary pass/fail against a fixed answer, gives Pattern Lab room to measure incremental improvement from a collaboration pattern even when no team clears the frontier outright, useful once MATH-500-style benchmarks stop discriminating between strong teams.
- notes
- Submitted 2026-09-21, within the discovery window. Distinct from math-500 (the only other tracked math-axis entry, saturated, fixed competition problems with symbolically-checkable gold answers): The Endless Exam instead uses parameterised, scalable construction families with an uncapped relative-quality score against a research frontier, explicitly designed to keep measuring progress once fixed-ceiling benchmarks like MATH-500 saturate. Verification marked det-gold: submissions are checked automatically for validity via compact certificates.
