SOTA2 Research · benchmarks

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
Object Hallucination Evaluation	POPE Random	multimodalcomputer vision	152	Jun 26, 2026
Visual Question Answering	GQA	multimodalcomputer vision+1	152	Jun 16, 2026
Logical Reasoning	Sudoku	nlp	152	Jun 19, 2026
Image-to-Text Retrieval	COCO	multimodalretrieval search rag