I don't notice a difference anymore
retained text
Am I the only one who basically doesn’t notice a difference between frontier models anymore? People say things like “Opus 5.5 is way better,” and I still use 5.5 high, but I genuinely don’t see much of a difference compared to other top models. Maybe my tasks just aren’t hard enough, but for normal use they all feel pretty similar to me. The biggest difference I notice is usually the product around them, like how Codex vs Claude Code formats responses, ... Are most model improvements mainly showing up on insanely hard benchmarks rather than normal everyday usage?
document fetched 2026-09-24 00:33 · canonical https://reddit.com/r/singularity/comments/1wodqq1/i_dont_notice_a_difference_anymore