SWE-Recreation - Nonfunctional Rate (NFR): leaderboard

SWE-Recreation (arXiv 2611.26311v1) software-recreation benchmark (Nonfunctional Rate (NFR)) grading coding agents rebuilding unfamiliar web/Android apps from pixels.

Metric: Nonfunctional Rate x100 (%, lower is better). Source: arxiv.org. Saturation forecast: Around February 2027. 5 models tracked. Updated 22 Nov 2026.

Top models

#ModelScoreOverall rank
1Kimi K324.9#62
2DeepSeek V4.1 Flash35.2#69
3GLM-5.3 Flash47.1#137
4GPT-6 Sol55.9#33
5Claude Opus 4.860.5#64

No result here: #2 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=swe-recreation-nonfunctional-rate-nfr · How It Works · Data refreshed daily, snapshot 2026-11-22.