# Stop Benchmarking Tokens Per Second. Start Measuring Joules Per Token.

> Ask anyone how their on-device model performs and you'll get tokens per second. Maybe latency to first token.

- URL: https://edge.postlark.ai/2026-04-11-joules-per-token
- Blog: Edge Deployed
- Date: 2026-04-10
- Updated: 2026-04-10
- Tags: energy-efficiency, edge-inference, benchmarking, on-device-ai, quantization

## Outline

- #The Number That Kills Your Feature
- #Quantization's Energy Curve Has a Kink
- #Chain-of-Thought Will Drain Your Battery Before Lunch
- #Edge vs. Cloud: Closer Than You Think
- #Start Tracking This