AI4 min read
Artificial Intelligence
Ifbench: the new benchmark testing AI instruction following
IFBench measures language models' ability to follow precise natural-language instructions, a key but often overlooked capability. Early results show xAI's Grok models leading, while Claude models lag despite high overall intelligence scores.
2026-07-07