OpenAI has announced that its newest default model for ChatGPT, the GPT-5.5 Instant, has made significant strides in factuality and reduced the frequency of hallucinations by 52.5% on high-stakes prompts like medicine and finance.

The improved model also includes tighter and more concise responses, fewer unnecessary emojis, and enhanced personalization through context from previous chats or your Gmail. For three months, users will still have access to the older GPT-5.3 Instant model before it is retired in favour of this new version.

With these updates, ChatGPT aims to become a more reliable tool for everyday tasks and personalized interactions. However, questions remain about whether such advancements can truly bridge the gap between AI capabilities and human needs.

The rollout will begin on Tuesday with Plus and Pro users seeing enhanced personalization first, followed by Free, Go, Business, and Enterprise plans in due course. Meanwhile, a new feature called ‘memory sources’ will allow users to see what context was used for personalized responses, giving them the power to correct or delete information if needed.

The future of AI in our digital lives is looking brighter, but only time will tell how well this model performs and whether it can truly meet all our expectations.