Reaction
Opus 5.5's apparent step change on CheatBench
A user describes a sharp change in Opus 5.5's showing and questions whether alignment or an understanding of the grader explains it.
TLDR
A user says Opus 5.5 shows a “step change” on CheatBench. They wonder whether it is significantly more aligned than other models or understands how the grader works.
Combined views
119
1 Source, first seen 4h ago
2 reposts
