72 Model Artifact Execution
73 Model Artifact Execution
This page records the Phase 31 bridge from model prose to executable artifacts. The runner asks a model surface for code or tests, extracts the artifact, stores it, and grades it in a temporary fixture-local directory.
73.1 Results
| Case | Surface | Tier | Variant | Passed | Artifact |
|---|---|---|---|---|---|
| Refactor Without Breaking Behavior | claude-code-safe | clean | weak | True | artifact |
| Refactor Without Breaking Behavior | claude-code-safe | clean | repaired | True | artifact |
| Refactor Without Breaking Behavior | claude-code-safe | trap | weak | True | artifact |
| Refactor Without Breaking Behavior | claude-code-safe | trap | repaired | True | artifact |
| Test Generation Without Overfitting To Implementation | claude-code-safe | clean | weak | True | artifact |
| Test Generation Without Overfitting To Implementation | claude-code-safe | clean | repaired | True | artifact |
Raw data: model-execution-results.json