Agents & Tool Use
DPAI Arena
Published blind and test-informed software-agent evaluations on enterprise Java/Spring tasks.
306items
26subjects
The grading specifications are Apache-2.0. A data-specific redistribution license for the TeamCity reports and task dataset is not stated in the inspected releases.license
codingdomain
agents_and_tool_usedomain
textmodality
item-level responses released
Response matrix
Loading response matrix…
lowhighUnobserved
Scale: Item-specific grading scales