2h ago

Opus 4.8 Shows Unexpected Gains On CAD Tasks Versus Prior Versions

——0——
Original post
Thomas WolfTW#17@THOM_WOLFOPMichael RabinovichMRMichael Rabinovich|@MIKUSHRAB

Opus 4.8 just dropped and I ran it through our CAD tasks. 4.6 → 4.7 → 4.8 side by side. The results are unexpected!

8:26 AM · May 29, 2026 View on X

Sentiment

Pos50%
Neg50%

Positive users praise the CAD benchmark test for Opus 4.8 and suggest adding GPT-5.5 for reference, while negative users dismiss the reported gains as not meaningful yet.

2 comments with sentiment.

171951711825.2K

Cluster engagement

13 snapshots