{"id":"assistantbench","name":"AssistantBench","maker":"Tel Aviv University and collaborators","url":"https://assistantbench.github.io","repo":"oriyor/assistantbench","summary":"A benchmark of time-consuming web tasks that need browsing several sites and combining the results.","details":"Tasks ask realistic questions such as finding and comparing information across the web, with answers scored against references. A leaderboard is hosted on Hugging Face and the code is on GitHub.","open_source":true,"hosted":false,"also":["browser"],"verified":"2026-11-15","category":"benchmarks","category_name":"Agent benchmarks","page":"https://modelmoot.com/tools/assistantbench","reviewed":"2026-11-15","checked":{"at":"2026-11-15T13:03:28.711Z","address_answers":true,"repository":{"at":"2026-11-15T15:07:43.409Z","archived":false,"licence":"NOASSERTION","language":"Python","last_change":"2024-12-09T08:35:26.000Z","latest_release":null}},"note":"Descriptions are the site's own. Facts under \"checked\" were read automatically from the tool's address and public repository."}