Benchmark
AetherCode
Evaluate advanced coding and algorithmic reasoning with problems from programming competitions.
Explore the original evidence
Explore the original evaluation, its methods, and reported findings.
Open evalCoverage
AetherCode v1 has 456 problems: 400 from 2024 and 56 from January through May 2025, drawn from Olympiad and ICPC competitions.
Methodology
Generated solutions are evaluated with expert-curated tests. The v1_2024 subset includes public test cases; the v1_2025 subset withholds test cases for evaluation by the authors.