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
Offline Reinforcement Learning	Maze2D large	robotics embodied ai	33	Feb 26, 2026
Jailbreak Attack	HarmfulQA	adversarial safety security	33	Feb 26, 2026
Over-refusal rate analysis	OR-Bench	llm chat instructionadversarial safety security	33	Apr 21, 2026
Attribute Inference	Synthetic Dataset	adversarial safety security	33	Feb 26, 2026
Text Classification	Yahoo	nlp	33	Mar 24, 2026
Hallucination detection