{"id":"inspect-evals","name":"Inspect Evals","maker":"UK AI Security Institute and contributors","url":"https://ukgovernmentbeis.github.io/inspect_evals","repo":"UKGovernmentBEIS/inspect_evals","package":"pypi:inspect-evals","summary":"A community collection of ready-made evaluations that run on the Inspect AI framework.","details":"Packages many benchmarks, including agent, coding, cybersecurity and reasoning evaluations, as Inspect tasks that run from a command line. The collection is open source under the MIT licence.","open_source":true,"hosted":false,"also":["observability"],"verified":"2026-11-15","category":"benchmarks","category_name":"Agent benchmarks","page":"https://modelmoot.com/tools/inspect-evals","reviewed":"2026-11-15","checked":{"at":"2026-11-15T14:03:44.997Z","address_answers":true,"repository":{"at":"2026-11-16T00:27:04.284Z","archived":false,"licence":"MIT","language":"Python","last_change":"2026-11-14T20:06:08.000Z","latest_release":{"tag":"v0.29.0","published":"2026-11-13T05:37:32.000Z"}}},"note":"Descriptions are the site's own. Facts under \"checked\" were read automatically from the tool's address and public repository."}