Jev vs GPT-5.6 Luna: 30 agent tool-routing decisions benchmarked
On the same synthetic cases, Jev routed with 100% accuracy at 323ms average latency versus 83.3% and 1821ms for GPT-5.6 Luna, prompting the author to build and open-source agent-tool-router.
Ran a small live benchmark for AI agent tool routing. Same synthetic cases, same tool registry. Jev vs GPT-5.6 Luna 30 routing decisions Jev: 100% accuracy 323ms avg latency 299ms p50 445ms p95 GPT-5.6 Luna: 83.3% accuracy 1821ms avg latency 1638ms p50 2577ms p95 Luna is a stronger general model, but for this narrow decision-layer task, Jev was faster and more consistent. That’s why I built van-vanich/agent-tool-router. npm: https://www.npmjs.com/package/@van-vanich/agent-tool-router GitHub: https://github.com/van-vanich/agent-tool-router
This post quotes another post: x.com/van_van1ch/status/2101984960684954030