ChatGPT 5.4 Pro Standard Mode - Adaptive Thinking or Nerfing Model?

It all just feels faster now. I think a lot of users probably noticed ChatGPT is way quicker lately, like maybe 4x faster or more.

Before, when I sent a big prompt, there was usually more time before it really got into Thinking, and then in the Thinking tab the steps would take longer. You could kind of feel it working through the task. Now it moves through everything much faster, almost like regular real-time chat.

For normal users, that probably feels better. Faster is usually good. But for long-context / high-effort work, I’m worried the speed is coming from less actual Thinking or earlier stopping. In my case, the faster run did not just finish faster — it missed the main deliverable.

I’m mostly trying to understand if this is the new normal, because then I need to know how to engineer around it, change my workflows, or look at alternatives.

I have Pro 20x, so I’m also wondering if maybe this is going to become a new subscription tier later, with fewer of these limits or deeper Thinking available.

Not saying I know what changed, because I can’t see the backend settings. But it feels similar to the recent Anthropic Adaptive Thinking change: everything got faster, but the deep reasoning feels like it may have been reduced. Curious if anyone else is seeing the same thing.