Benchmark

A benchmark is a standard test used to measure and compare how well an AI system performs on a specific task.

A benchmark is a set of test questions or tasks with known answers or scoring rules. It gives people a common way to compare AI systems and check whether changes improve or weaken performance. For everyday AI tool users, benchmarks are useful context, not a guarantee. They can show which tool may handle a type of work better, but they do not prove it will work well on your own task. Example: a benchmark for summarizing might ask an AI to shorten several sample articles, then compare its summaries with reference summaries using clear scoring rules.