- ChatGPT default model has been updated to GPT-5.3 Instant to make conversations easier
- OpenAI promises this will “reduce the embarrassment” of warnings and rejections
- ChatGPT 5.3 should also have fewer hallucinations
OpenAI has heard complaints that ChatGPT too often answers a simple question with a long list of limitations, and has released GPT-5.3 Instant as the new default model for ChatGPT to, in its words, “reduce the embarrassment” of those difficult conversations. The update is designed to reduce unnecessary rejections and shorten security disclaimers to improve the overall tone of AI.
Previous versions of the system sometimes took a defensive approach when answering questions. Even when a request was harmless, the model might begin with a lengthy explanation of safety rules before finally addressing the prompt. In other cases, you might refuse to answer a question entirely because you misunderstood the intent behind it.
GPT-5.3 Instant attempts to correct that behavior. Instead of starting with warnings or hypothetical misuse scenarios, the system now prioritizes providing a direct answer when the question is reasonable. Safety restrictions remain in place, but the model has been adjusted to avoid unnecessary precautions.
The update will simply appear the next time most people open the chatbot and should boost ChatGPT’s intellect as well as its social niceties. It’s supposed to enhance the subtler qualities that determine whether talking to an AI is pleasant or annoying: the “chill,” so to speak.
We hear your feedback loud and clear, and 5.3 Instant reduces embarrassment. pic.twitter.com/WqO0XzLcVuMarch 3, 2026
Now, if I were young and hip, saying something that “reduces shame” would only make me think that the shame factor was even greater. But if the new model means ChatGPT is less rigid and doesn’t implement a series of warnings with every message, I’ll accept the slightly desperate wording that OpenAI uses.
There should also be less to warn about if you have the improved accuracy promised by OpenAI. ChatGPT-5.3 Instant should have fewer hallucinations – about 27% less when doing online research, though only 20% less without going online to check. You should also be better at combining web data with your existing knowledge.
GPT-5.3 Personality
Replacing ChatGPT 5.2 with ChatGPT 5.3 changes the core experience across millions of daily chats. The update covers everything from homework help to recipe ideas to casual curiosity about how something works. GPT-5.2 Instant will not disappear completely. Paid users will still have access to it as a legacy model.
The emphasis on soft skills is a notable evolution from when every AI developer boasted about solving the toughest problems or performing the most complicated tasks. The focus is now on a product that is pleasant to use, and that means making AI pleasant to talk to.
A chatbot that sounds overly cautious is irritating, and if you want to keep people coming back, even subtle design decisions about response style can be crucial. The winner of the AI chatbot battle may not be the flashiest or most powerful model, but one that doesn’t remind people of an annoying pedant who takes five minutes to answer a question when a single sentence will do.
And of course you can also follow TechRadar on TikTok for news, reviews, unboxings in video form and receive regular updates from us on WhatsApp also.
The best business laptops for every budget
