If you like ChatGPT, it’s just been updated.
OpenAI on Thursday introduced GPT-5, which the company says offers both a “significant leap in intelligence as well as fewer hallucinations. The new protocol is said to be better at following instructions and minimizing sycophancy.
As such, OpenAI has stated that the new models offers improved performance across coding, math, writing, health, visual perception, and more. Other improvements include handling of complex front-end generation and debugging large code repositories.
Other improvements include enhanced writing capabilities and improved ability to answer health-related questions.
According to OpenAI, GPT-5 is “significantly less likely to hallucinate” compared to prior models. With web search enabled, GPT-5’s responses are 45 percent less likely to include a factual error than GPT-4o, and when thinking, responses are 80 percent less likely to have errors than OpenAI o3.
The GPT-5 protocol is less likely to lie to the user about tasks it is able to complete or what has been done, and more honestly communicates its actions and capabilities for tasks that are impossible, underspecified, or missing key tools.
Where ChatGPT is concerned, users can select from four personalities, including Cynic, Robot, Listener, or Nerd. The different personalities are opt-in and let users set how ChatGPT interacts and responds to questions.
The GPT-5 protocol is available to all ChatGPT users, although Plus subscribers will get more usage before a limit is imposed, and Pro subscribers can access GPT-5 Pro, a version that has extended reasoning.
When users hit a usage limit, ChatGPT will switch to a mini version of GPT-5 for remaining queries. GPT-5 replaces GPT-4o, OpenAI o3, OpenAI o4-mini, GPT-4.1, and GPT-4.5.
If you’ve had a chance to try the updated ChatGPT features, please let us know about your experience in the comments.




