Technology ❯ Artificial Intelligence ❯ Performance Evaluation

Benchmarks

STEM Competitions Terminal-Bench 2.0 Industry Records Complex Agent Capabilities SOTA Models IFEval Comparative Analysis GSM8K and MATH500 AI Benchmarks SWE-bench

New HANRAG Preprint Targets Multi‑hop Noise as RAG Guides Proliferate

The heuristic framework proposes query decomposition and noise filtering to boost QA performance, with findings reported as a preprint.