Felan AI ← All Extension Results

Extension Benchmark Result

Codebase Memory

A three-case Codebase Memory comparison using the ratio of median-reduced case costs.

Cost vs baseline 27.8% lower $5.7271 baseline → $4.1357 candidate
Agent-step duration 7.8% shorter 1,167,282 ms baseline → 1,076,719 ms candidate
Framework pass rate 100% 18 of 18 published attempts passed

Cost by case

Case Baseline median Candidate median Case gain
Extension architecture$1.2002$1.21251.0% higher
Extension config scope$3.8948$2.243242.4% lower
Project instructions$0.6322$0.68007.6% higher

Quality and evidence notes