11 points spenvo 1 day ago 6 comments
simianwords 1 day ago | parent
It’s better than both fable and Astra? I’m going to start trusting this index less and less.
ramon156 1 day ago | parent
the issue is, what will we use? companies are benchmaxxing, so how do you make a bench that cannot be cheated?
Human curated benches aren't accurate enough
epolanski 1 day ago | parent
The problem with benchmarks is that they are end to end.
You go from one prompt to the final solution, whereas for many of us it is about the experience of iterative, multi turn working.
itzikkatz 1 day ago | parent
If its price is almost the same as Opus 5.5 but it’s ultimately not as good, I don’t understand the use case for this model. Perhaps Haiku 5.5 could fill a real need.