Of course, this is based on the standard when Fable was nerfed and the debuff was applied that made it constantly switch to Opus 4.8.
Even if things were going well, once it switched to Opus 4.8, I felt like the long work context was broken, so my satisfaction level wasn't high.
On the other hand, GPT 5.6 Sol feels very solid in handling long tasks (paper assistance) without errors, so my satisfaction level is high.
To put it simply, after Fable was nerfed and switched to Opus, I double-checked the research papers I had verified with 5.6 Sol.
It caught two errors that I had missed for about a month.
Based on this, when I questioned Opus 4.8, it coolly admitted its mistake. Jerk...
