A New Frontier-Level Math Benchmark

Nova-Math is a government-supported project (Korea’s Ministry of Science and ICT (MSIT) and the National Information Society Agency (NIA)) to build a calibrated mathematics benchmark that evaluates and challenges large language models (LLMs). We are recruiting problem writers to create original, high-quality items across four difficulty tiers: Challenge, Hard, Medium, and Easy. All accepted items are paid on a per-item basis, and authors will be included in a public paper describing the dataset, evaluation protocol, and baselines by year-end.

Instructions for Contributing

  • Consult the FAQ document and read all instructions.
  • Write a novel question.
  • Check your question. Do not submit questions to AI chat models. To check your question's difficulty level, use our tool.
  • Submit your question using this form.
  • After submission, your question will be checked for difficulty and uniqueness. If the question is accepted, you will receive an NDA and IP transfer document. After that, you will be compensated.

Compensation

Update
We have finished collecting Challenge level questions. Submissions for this tier are closed.
Tier Mathematical Depth Explanation Compensation (per question)
Challenge Graduate Research Not solvable by current AI; typically only a few human experts in related areas can solve. A good Challenge problem has a novel core insight and resists back solving, guessing, or relying on heuristics. $3,623
Hard Advanced undergraduate or beginning graduate level
(e.g., Real/Complex Analysis, Algebra, Probability, Combinatorics, Geometry, Topology)
Demands sustained reasoning or creative application of known results. A “Hard” problem may resemble a strong qualifying-exam question or a compact research-style exercise that exposes a subtle mathematical structure or method. $350
Medium Standard undergraduate level
(e.g., Calculus, Linear Algebra, Abstract Algebra, Discrete Math, Elementary Probability)
Tests conceptual understanding and multi-step reasoning rather than routine computation. A good “Medium” question highlights an important idea that generalizes or connects topics, without excessive technical detail. $72
Easy Early undergraduate or high school enrichment level
(e.g., Precalculus, Elementary Number Theory, Geometry, Basic Statistics)
Focuses on clarity, correctness, and engagement. An “Easy” question illustrates a fundamental concept cleanly, with an elegant or surprising twist accessible to a broad audience. $36

Important Dates

  • September 1, 2025: Submissions open
  • November 26, 2025: Submissions close
  • December 2025: Competition concludes and report is released.

Contact

For questions about the benchmark or to discuss collaboration opportunities, please contact novamath.faq@gmail.com.