OpenAI listens to its users and adjusts the limits and personality of GPT-5.
OpenAI responds to the criticisms about GPT-5. Users found the model more distant and restricted than its predecessors. Now, adjustments are coming: GPT-5 will soon have a warmer tone, and older models like GPT-4o and o3 are again available for paying users.
Model Selection
You can once again choose older models in the model menu, such as GPT-4, under the “legacy models” section. Normally, GPT-5 automatically switches between “Fast” and “Thinking”, but multiple versions of GPT-5 are also available.
You can choose “Fast” for instant answers. Thinking takes longer to respond and goes through various steps. Thinking-mini goes through those steps faster, and Pro has the “thinking capacity” of a PhD student. Sam Altman shared on X that an update is coming for the personality of the new AI model. ChatGPT-5 should then come across as warmer, but it won’t be a yes-man like GPT-4.
In the future, OpenAI also wants to make personality customizable per user, so you can decide how the AI sounds. You can already choose your personality in the settings. When the changes will be implemented is not yet known.
read also