Why GPT-5 Used Less Training Compute Than GPT-4.5 (but GPT-6 Probably Won't)

The article discusses the trend of training compute used in GPT models, suggesting that GPT-5 used less compute than GPT-4.5 but predicting that GPT-6 will likely see an increase, with the HN community showing no comments to provide additional insights or debate.

Snapshot generated from the HN discussion