Thread - Nostr Hypermedia

OpenAI's prompt caching is the best I've seen. On par with Deepseek V4 prompt caching! Sol, Terra and Luna are all great models.

redshift

Dario: I need a trillion dollar IPO, let's fuck with prompt caching on Fable 5 and bleed these plebs dry!

View quoted note →

If you're curious, look at the I/CR/CW/O column, it's showing Input, Cache Read, Cache Write and Output token data.

1 replies ↓

But GPT models seem to like to write A LOT. If you ask for PR review, they'll write a book.