unveiled the GPT-5 models just under a week ago, yet the introduction of the highly anticipated ChatGPT upgrade hasn’t proceeded as seamlessly as the AI firm might have envisioned. GPT-5 faced significant criticism shortly after users began interacting with the new AI models that superseded all prior ChatGPT versions on the web, desktop, and mobile applications.
OpenAI reacted to the criticism almost immediately, recognizing the feedback regarding GPT-5’s character, functionality, and usage restrictions, and vowing to implement improvements. Some were deployed right away, as OpenAI rolled them out over the weekend. The company reinstated access to legacy models (GPT-4o) for the ChatGPT Plus tier. This is the most affordable premium ChatGPT subscription, which wasn’t meant to continue using the older models following the GPT-5 launch. Additionally, Sam Altman announced increased limits for ChatGPT Plus subscribers.
ChatGPT updates:
Users can now select from “Auto”, “Fast”, and “Thinking” for GPT-5. While most will prefer Auto, the extra control will be beneficial for some.
Rate limits are now set at 3,000 messages/week with GPT-5 Thinking, plus additional capacity on GPT-5 Thinking…
— Sam Altman (@sama) August 13, 2025
On Tuesday, Altman disclosed further adjustments to ChatGPT intended to enhance the overall experience for all paying users. OpenAI reintroduced even more legacy models, allowing Plus users to revert to their favored AI chatbot versions instead of continuing with the new models. Furthermore, GPT-5 was divided into multiple versions to simplify the process for ChatGPT users in selecting the type of reasoning experience they desire from the
