Skip to content

Comment on Prompt caching for cheaper LLM tokensparent

Comments

Really well done article.

I'd note, when I gave the input/output screenshot to ChatGPT 5.2 it failed on it (with lots of colorful chain of thought), though Gemini got it right away.

Huh, when I was writing the article it was GPT-5.1 and I remember it got it no problem.

AboutSource Built by g1lg1l

Hackerly is an independent reader for Hacker News, built on the public HN API. Not affiliated with Y Combinator.