logoalt Hacker News

lsllcyesterday at 9:38 PM2 repliesview on HN

Interesting that Sol (low) did better than Sol (medium) in your benchmark (and is barely more expensive than Terra). I too have been using 5.3 codex as a cheap-but-good model and are switching to Terra (xhigh).


Replies

XCSmeyesterday at 9:44 PM

Yeah, for some reason the (low) versions do really well, like they think directly of the solution instead of going around all the edge-cases and getting lost in one of them.

XCSmeyesterday at 9:44 PM

Also for most, there doesn't seem to be a big difference between (medium) and (high).