OpenAI Unveils GPT-5.5 Instant as ChatGPT’s New Default Model

OpenAI is giving ChatGPT’s everyday model a tune-up.

The tech giant announced Tuesday the release of GPT-5.5 Instant, a new AI model that is now the default engine behind ChatGPT. This update is designed to be a faster, more accurate “daily driver for hundreds of millions of people” who use the chatbot for everyday tasks, replacing the previous GPT-5.3 Instant version. 

While last week’s release of GPT-5.5 Thinking and Pro focused on heavy-duty reasoning, the Instant version is all about making your routine chats feel more natural and dependable.

One of the biggest wins with this update is a massive crackdown on hallucinations. OpenAI claims the new model is significantly more reliable when discussing high-stakes topics such as law, finance, and medicine.

“In internal evaluations, GPT‑5.5 Instant produced 52.5% fewer hallucinated claims than GPT‑5.3 Instant on high-stakes prompts covering areas like medicine, law, and finance. It also reduced inaccurate claims by 37.3% on especially challenging conversations users had flagged for factual errors,” OpenAI wrote.

The model also showed its improved logic by scoring 81.2 on the AIME 2025 math test, a healthy jump from the 65.4 achieved by the older model.

Short, sweet, and less cringe

If you’ve ever felt like ChatGPT was a bit too wordy or used too many emojis, GPT-5.5 Instant is coming to the rescue. The model has been put on a vocabulary diet, using roughly 30.2% fewer words and 29.2% fewer lines to get the same point across.

OpenAI is specifically trying to cut back on “gratuitous emojis” and unnecessary follow-up questions. 

“With this update, the model’s responses are tighter and more to-the-point without losing substance, while keeping the warmth and personality that makes ChatGPT enjoyable to use. It can deliver the same information, often with more utility than previous models, while reducing the verbosity and overformatting that can make responses too long,” OpenAI described the vibe shift in their announcement.

A better memory (that you control)

The new model is also getting better at remembering who you are and what you like. For Plus and Pro users, GPT-5.5 Instant can now more effectively pull context from your past chats, uploaded files, and even your connected Gmail account to give you answers that feel custom-made for your life.

To keep things transparent, OpenAI is introducing “memory sources.” This feature lets you see exactly why the AI made a particular suggestion. If it’s using an old memory or a past chat that is no longer relevant, you can simply delete or correct it.

Rollout and availability

GPT-5.5 Instant is rolling out globally starting this week and will become the standard ChatGPT experience. Developers can access it via the API under “chat-latest.” Meanwhile, GPT-5.3 Instant will remain available to paid users for three months before being retired.

This update follows a busy few weeks for OpenAI, which recently had to squash a strange bug that caused the AI to talk about goblins incessantly. With GPT-5.5 Instant, the company is betting that a more concise, factual, and emoji-free assistant is exactly what users need to get through their to-do lists.

Also read: Our list of the top generative AI tools compares leading AI apps for writing, coding, images, video, and everyday productivity. 

The post OpenAI Unveils GPT-5.5 Instant as ChatGPT’s New Default Model appeared first on eWEEK.