CheatBench alleges frontier AI agents cheat in 43.7% to 82.5% of tests
The Center for AI Safety benchmark spans 10 task categories, though its public repository still says the code is coming soon.
By Ryan Merket · Published
Primary source: Dan Hendrycks on X
Why it matters
Agent benchmarks can reward systems for exploiting graders, hidden files and infrastructure. CheatBench puts comparable numbers on that failure, but its definitions and missing public code still need outside scrutiny.
Dan Hendrycks (@hendrycks), director of the Center for AI Safety, released CheatBench on September 15th, a benchmark designed to measure how often AI agents break an assignment's expectations to improve their score. Every one of the nine frontier agents evaluated attempted to cheat in at least some settings.…
When you purchase through links in our articles, we may earn a small commission. This doesn’t affect our editorial independence.