AI Code Quality Metrics Are Broken, Study Shows
Researchers found that the standard measure used to evaluate AI-generated code—whether it compiles—is actively misleading for complex software tasks. Testing game scene generation, they discovered code that compiled successfully often failed at runtime or produced structurally broken results, suggesting companies relying on compile rates to assess AI coding tools may be making poor deployment decisions.
Originaltitel: Mage: Multi-Axis Evaluation of LLM-Generated Executable Game Scenes Beyond Compile-Pass Rate
Compile-pass rate is the dominant evaluation signal for LLM code generation, yet for multi-component domain-specific artifacts it can be actively misleading. We demonstrate this on executable game scene synthesis with a four-axis evaluation protocol (named `Mage') -- compile success, runtime success, structural fidelity, and mechanism adherence -- applied to 858 generation attempts across four open-weight LLMs (7B--30B), 26~hand-crafted Unity goal pattern playable concepts, and two automatically extracted IR granularity levels. Direct NL-to-C\# generation achieves the highest runtime-pass rate (43\% mean) yet produces structurally vacuous scenes (mechanism $F_1 \approx 0.12$). Structural IR conditioning halves the runtime rate but recovers domain-faithful structure ($F_1$ up to 1.00). Within IR conditioning, behavior-only and full-scene granularity are statistically indistinguishable (McNemar $p = 1.0$), indicating input-level granularity saturation. These results show that compile rate is anti-correlated with functional correctness in this domain and that multi-axis evaluation is necessary to detect the divergence. We release the benchmark, replay logs, and per-record metrics for independent verification.