Benchmarks

Compare state-of-the-art methods across the tasks and datasets used to measure AI progress.

309,131 indexed

Task	Dataset	Domain	Results	Last Update
Image-Text Alignment Evaluation	nocaps out-of-domain (val)	computer visionnlp+1	40	May 1, 2026
Untargeted Adversarial Attack	ImageNet	adversarial safety security	40	May 1, 2026
Out-of-Distribution Detection	CIFAR-100 ID Near-OOD Average OpenOOD v1.5	computer visiontabular time series	40	May 27, 2026
Math	Minerva	llm chat instruction	40	May 28, 2026
Question Selection and Generation	America in One Room: Democratic Reform (A1R)	nlp