Blender Agent Benchmark

Measure changes with the same tasks, model, effort, limits, Blender build, and evaluator. Preserve natural agent behavior.

Protect benchmark integrity

1. Create isolated directories for every condition and repetition.
2. Do not leave the other condition's code, renders, metrics, or expected fixes where the agent can discover them.
3. Keep the user-facing task prompt identical except for explicit skill invocation in the plugin condition.
4. Use codex exec --ignore-user-config for the no-plugin baseline.
5. Use the installed plugin in a fresh invocation for the plugin condition.
6. Record CLI version, model, effort, Blender build, duration, tool calls, failures, and output hashes.
7. Evaluate outputs after generation. Do not leak hidden rubric details to the agent.