I wouldn't say vice-versa is true. GPT 5.5 routinely finds major mistakes made by Opus 4.7, but I've yet to have it work the other way around.

Additionally running GPT-5.5 on medium sometimes gives me better results than high mode. On any of them I still have to push the models in the right direction.