AI Paper Atlas

Research map / Benchmarks

Benchmark leaderboards

1,532 benchmarks and 2,368 leaderboards from the results tables of 116,039 papers at ICML, NeurIPS, ICLR, CVPR, AAAI, ICCV and ECCV, and the journals TPAMI, IJCV and JMLR (2016–2026). For each benchmark, every paper's best reported number for its own method, ranked, with the table it comes from.

Numbers are as each paper reports them, read by AI from its tables and not checked by hand. Papers on one benchmark can differ in setting (training data, model size, extra data, split), so the topic of each paper is shown next to it.