Mefa @Mefa_MC - Twitter Profile

Codex's latest limit reduction makes it almost unusable I’ve been using Codex regularly for around eight months, and it has worked extremely well for my projects. The limits used to be generous enough to experiment, make changes, correct mistakes, and still get meaningful work done. I understood the earlier reductions and kept hoping the limits might improve again. Even after those reductions, Codex was still somewhat usable. But after the latest reset, the limits feel far too restrictive. I have almost completely stopped using Sol because, even at medium, a single task can sometimes consume 5–10% of my weekly usage. It can do good work, but the usage cost is now so high that I avoid it. I mostly use Terra or Luna at medium, and even then I think several times before asking Codex to do anything. One misunderstood instruction can waste a significant part of the allowance, and correcting the mistake consumes even more. I’m a developer, so I can do many tasks myself and only use Codex for work that would otherwise take a long time. But for non-coders, or people with limited coding knowledge who depend heavily on Codex to build something, the current limits seem completely impractical. I understand that these models are expensive to operate. I would honestly prefer paying a little more and getting the previous usable limits back rather than paying less for something I’m afraid to use. TL;DR: The latest Codex limits are so restrictive that I now hesitate before every request. Sol can consume 5–10% of my weekly usage on a single task, while even Terra and Luna feel heavily limited. I would rather pay more for genuinely usable limits. What is everyone else planning to do? Are you hoping the limits improve, staying with Codex, switching to another service, or considering a local setup?

30

204

9

8

9K