ESxqouUav
@ywbbonvdz
Claude 3.5 is still outperforming 3.7 at writing Convex code! Both are still #1 and #2 though. We've found GPT-4o less responsive to prompting. We maintain an active leaderboard on our website showing model performance on a standard set of coding tasks. https://t.co/7xE5rso4gh
0 reply
0 recast
0 reaction