SmellBench: leaderboard
Architectural code smells erode software maintainability and are costly to repair manually, yet unlike localized bugs, they require cross-module reasoning about design intent that challenges both developers and automated tools. While large language model agents excel at bug fixing and code-level refactoring, their ability to repair architectural code smells remains unexplored. We present the first
Metric: Weighted Effectiveness (E) (self-reported). Source: benchmarklist.com. Saturation forecast: Around July 2027. 10 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.3 Codex | 44.4 | #69 |
| 2 | Gemini 3.1 Pro (Preview) | 43.5 | #58 |
| 3 | GPT-5.4 | 42.8 | #79 |
| 4 | Claude Haiku 4.5 | 37.9 | #280 |
| 5 | Claude Opus 4.6 | 37 | #63 |
| 6 | Claude Sonnet 4.6 | 35.4 | #90 |
| 7 | GPT-5.1 Codex Max | 34.7 | #100 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=smellbench · How It Works · Data refreshed daily, snapshot 2026-10-20.