Opus is the best coding model
On ProgramBench as tested by Vals AI it almost resolves 15.5% of problems while GPT-5.5-xhigh is at 9.5%.
7:50 PM · Jun 2, 2026 · 31.2K Views
Opus Leads ProgramBench With 15.5% Resolution Rate Over GPT-5.5
--0--
Sentiment
Sentiment building, check back later.
Cluster Engagement
-
Views
-
Comments
-
Reposts
-
Bookmarks
Expand data
Posts from X
Most Activity
Most ActivityTimeline
VIEWS2.3KBOOKMARKS1LIKES29REPLIES1
Lisan al Gaib@scaling01
costs are of course horrendous which is my permanent underclass ass had to buy OpenAI subscription
Lisan al Gaib@scaling01
Opus is the best coding model
On ProgramBench as tested by Vals AI it almost resolves 15.5% of problems while GPT-5.5-xhigh is at 9.5%.
12hViews 2.3KLikes 29Bookmarks 1