OpenAI is updating ChatGPT with a new default model, called GPT-5.5 Instant, which the company says starts rolling out today. Multiple outlets report that the change is intended to make ChatGPT’s answers feel “smarter” and more accurate. OpenAI also says the upgrade reduces what it describes as “gratuitous emojis,” aiming to lessen emoji-heavy or unnecessary additions in responses.
CNET and 9to5Mac both focus on the company’s stated goal of cutting down on unsolicited emoji use, describing it as part of how the new model changes output. TechRadar adds additional reported user-facing effects, including that responses may be shorter and that users may notice fewer hallucination-like issues. Across the coverage, the common theme is a backend model swap for ChatGPT that adjusts how the system generates responses, with improvements highlighted around accuracy and a cleaner tone.
The outlets do not provide independent measurements, but they consistently attribute the improvements to OpenAI’s announced behavior changes tied to GPT-5.5 Instant.
