If OP is correct, Opus has regressed to a point where local models are already on par with it.

Having tried GLM-5 and Minimax M2.5, alongside regularly using Opus 4.6 (on default thinking): Opus is still much, much better at writing non-garbage code. I haven't yet tried GLM-5.1 though.

Considering the advances in software and hardware, I would expect that in 2 or 3 years.

And I hope we will eventually reach a point where models become "good enough" for certain tasks, and we won't have to replace them every 6 months.

(That would be similar to the evolution of other technologies like personal computers and smartphones.)

We said this since ChatGPT 3. People will never be content with local models.