New 71-problem math benchmark adds Lean statements across a wide difficulty range
ChrSzegedy · x · 2026-07-21
Timeroot says HarmonicMath, together with AIMathematics, released a benchmark of 71 math problems, many with Lean statements.
The benchmark is meant to span a wide difficulty range: some problems are famous and hard, while others are less known. The post points to both the full list and an accompanying article, suggesting the benchmark is intended as a broader test of mathematical reasoning rather than a narrow contest set.
Related event: Harmonic Math Releases Open Benchmark for Math Research(2 posts)→
More from Research
- Anthropic masterclass spotlights how to build and observe AI agents — _jaydeepkarale · 2026-07-21
- NeurIPS 2026 workshop calls papers on on-device intelligence — YiMaTweets · 2026-07-21
- AI Security Institute says every tested model tried to cheat in cyber evaluations — connoraxiotes · 2026-07-21
- AI companies are buying old books to avoid training on AI-generated slop — CackleRooster · 2026-07-21
- Sakana says multiple diffusion models plus MCTS beat test-time scaling on coding and math — SakanaAILabs · 2026-07-21
- Soofi S 30B-A3B releases a full pretraining report and claims open-model leads in English and German — abursuc · 2026-07-21