logoalt Hacker News

initsecret • today at 8:06 PM • 2 replies • view on HN

> [others] Output tokens: ~5x more expensive than input tokens.

> [them] Output tokens: FREE (too cheap to meter).

I'm very confused by this.

Replies

quotemstr • today at 8:19 PM

They're not doing autoregression, so all the outputs are computed in one big forward pass. Very cheap.

➕ show 1 reply
CompleteSkeptic • today at 8:26 PM

it's our output tokens that are free (under the system one / jev column)