Simon Willison(@simonw):GPT-5.6 found optimizations that "reduced end-to-end serving costs by 20%" for OpenAI to serve that model Presumably that's billions of dollars a month in savings at this point?

2026.07.30 01:34

GPT-5.6 found optimizations that "reduced end-to-end serving costs by 20%" for OpenAI to serve that model Presumably that's billions of dollars a month in savings at this point?

Vaibhav (VB) Srivastav@reach_vb

2026.07.29 21:38

Codex analysed production traffic, improved load balancing, rewrote production GPU kernels and ran hundreds of experiments on its own speculative-decoding model. The kernel improvements reduced end-to-end serving costs by 20%, while speculative decoding improved token-generation efficiency by more than 15%.