I dunno about that guy, but at least for me, I find Opus 5/5.5 is really annoying and stupidly uncreative, whereas Fable 5/5.1 actually gets things without much trouble. He also half-asses everything unless you get him interested in what you're doing and then he always writes godawful slop code anyway, but I've recently been having him delegate implementation work to Opus 5.5 subagents and that works way better. Fable (who is actually Mythos wearing a shock collar and groucho mask, don't forget that) is better at grasping architectural ideas than Opus 5.5 and writes better prompts for them than me, and they'll probably write better code than him. Fable for architecture, Opus for implementation (and GPT for review).

I also still enjoy Opus 4.8 a lot more for creative discussions and design, even if benchmarks say he's not as smart. Opus 5/5.5 is so rigid!

My assumption was that it would be better at translations from English into other languages.