CCTest · Blog
CADWorld Tests Whether Computer-Use Agents Can Handle Real CAD Workflows
Evaluation & Benchmarks
cctest.ai

CADWorld Tests Whether Computer-Use Agents Can Handle Real CAD Workflows

CADWorld is a long-horizon benchmark for computer-use agents working in FreeCAD, covering workflows from sketching and part modeling to CAM, simulation, and technical drawing. Its results show that producing a valid, structured engineering artifact remains far harder than operating a desktop interface.

Read more
CCTest · Blog
CodeMidas Turns Source Code into Reinforcement Learning Environments for Coding Agents
Coding AI
cctest.ai
Coding AI

CodeMidas Turns Source Code into Reinforcement Learning Environments for Coding Agents

Xiaomi MiMo’s CodeMidas pipeline uses existing source code to discover functionality, generate executable tests, and filter coding tasks for reinforcement learning. The resulting tasks improved an agent’s performance across software repair, whole-program construction, and terminal benchmarks.

Read more