Every story and comment from the Hacker News API, in a layout borrowed from 2001.
A comment on System Card: Claude Mythos Preview [pdf]. Parent comment
Because 100% is maximum, you should be looking at error rates instead. GPT has 25% on Terminal Bench and the new model has 18%, almost 1.4x reduction.
There are no comments on this comment yet.
Because 100% is maximum, you should be looking at error rates instead. GPT has 25% on Terminal Bench and the new model has 18%, almost 1.4x reduction.