ChatGPT Plus Brings Back Old Models With Updated GPT-5 Experience

Sam Altman returns the old model in ChatGPT (photo: x @ivnways)

OpenaI introduced a new feature that allows ChatGPT Plus users to choose from various GPT-5 modes, as well as restore popular previous AI models.

On August 13, 2025, Andrew Grush reported that OpenAI had responded to criticisms aimed at its latest model, the GPT-5. After many users, including a reviewer such as Calvin Wankhede of the Android Authority, complained about the GPT-5 that felt too short and lacked "private," the company quickly made improvements to the performance, character, and constraints on the use of the model. Recent updates are now making user experience even better, especially for ChatGPT Plus customers.

OpenAI is expanding the GPT-5 functionality by adding several new modes: Auto, Fast, Thinking Mini, and Thinking Pro. According to GPT-5 in certain scenarios feels superior to other alternatives, and it personally likes its concise and grounded tone, which is considered more natural than previous models, GPT-4o.

However, realizing that not everyone has the same preferences, OpenAI is taking steps that many parties appreciate. They are now providing back some of the old models that users miss a lot. These models include:

GPT-4o: The previous standard model, designed for general use.

GPT-4.1: A slightly better model for specific and detailed questions.

o3: The model was previously the main choice for deep questions and philosophical reflections.

o4-mini: The "thinking" model is similar to the o3, but more suitable for simple questions.

This move is considered as proof that OpenAI listens to input from their customers. By providing more models and flexibility options, the company is expected to continue to improve their services over time.