[trash_cat]

I think the naming schemes are quite arbitrary at this point. Going to 5 would come with massive expectations that wouldn't meet reality.

[mrandish]

After the negative reactions to GPT 5, we may see model versioning that asymptotically approaches the next whole number without ever reaching it. "New for 2030: Claude 4.9.2!"

[esafak]

Or approaching a magic number like e (Metafont) or π (TeX).

[Squarex]

the standard used to be that major version means a new base model / full retrain... but now it is arbitrary i guess