Very (reasoning) heavy benchmarks do seem like the way to go, being the hardest ... | Hacker News

Hacker Newsnew | past | comments | ask | show | jobs | submit

alansaber 4 months ago | parent | context | favorite | on: SWE-bench Verified no longer measures frontier cod...

Very (reasoning) heavy benchmarks do seem like the way to go, being the hardest to game.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact