Y
Hacker News
new
|
ask
|
show
|
jobs
by
AS04
317 days ago
400k context with 100% on the fiction livebench would make GPT-5 the undisputably best model IMHO. Don't think it will achieve that though, sadly.