ROUTEBusiness workflows
JEV Decision API Benchmark
Alp@alperenerol
This project benchmarks the typesafe/jev-1.13 decision API on 34 labeled triage cases, demonstrating its performance in choice routing and urgency gating.
Benchmarked typesafe/jev-1.13 on 34 labeled triage cases:
choice routing 14/14, urgency gating with 0 false positives, score MAE 0.18, ±0.004 drift across runs, ~$0.016 per 1k decisions.
Not a chat model — a typed decision API. Dataset + harness: github.com/alperenerol/jev-1.13-mini-bench