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
Classification	Ten-dimensional normal data (d=10) (test)	computer visionnlp+3	30	Jun 23, 2026
System Prompt Corruption (SPC)	COMA Evaluation Suite	adversarial safety security	30	Jun 23, 2026
Question Answering (Q&A)	COMA Evaluation Suite	nlp	30	Jun 23, 2026
Adversarial Task Switching (ATS)	COMA Evaluation Suite	adversarial safety security	30	Jun 23, 2026
Image Forgery Localization	Columbia Unseen Domain	computer vision