The China Academy of Information and Communications Technology (CAICT) has
opened a public solicitation for test tasks for its Fangsheng AI agent
benchmark. The benchmark uses a layered testing framework covering basic
capabilities, representative scenarios and complex integrated tasks, and aims to
extend testing to comprehensive capability measurement, process analysis and
problem diagnosis. CAICT is seeking high-quality task submissions to refine test
tasks, testing environments and evaluation methods and to boost the benchmark’s
professionalism, representativeness and industry applicability.
