OpenAI Overhauls GPT-5 with Modes, Limits, and Personality Upgrade
- GPT-5 now offers Auto, Fast, and Thinking options for users.
- Rate limits raised to 3,000 messages per week for GPT-5 Thinking.
- Model to get a warmer, customisable personality.
OpenAI has announced a series of updates to its flagship AI model, GPT-5, following user backlash. CEO Sam Altman revealed the changes in a post on X, highlighting new usage modes, increased rate limits, and a refreshed personality for the model.
ChatGPT users will now be able to select between Auto, Fast, and Thinking modes while working with GPT-5. The change follows growing demand for advanced reasoning capabilities, with Altman noting that daily usage of reasoning models has risen from less than 1% to 7% among free users and from 7% to 24% among Plus subscribers.
To accommodate this growth, OpenAI is raising rate limits to 3,000 messages per week for GPT-5 Thinking, with extra capacity on GPT-5 Thinking mini. Altman cautioned that limits may be adjusted in the future based on usage trends.
Paid ChatGPT users will also gain a 'Show additional models' option in settings, enabling access to models such as o3, 4.1, and GPT-5 Thinking mini. However, GPT-4.5 will remain exclusive to Pro subscribers due to its high GPU cost. Altman also confirmed that GPT-4o has been restored to the model picker, reassuring users that any deprecation would come with prior notice.
Another major update is in GPT-5’s personality. OpenAI is working on making it 'warmer than the current personality but not as annoying (to most users) as GPT-4o', with a long-term goal of enabling per-user customisation for tone and style.
These updates aim to address user concerns while enhancing flexibility and personalisation, marking a significant step in GPT-5’s evolution as OpenAI continues to balance performance, usability, and resource costs.
