OpenAI has updated the default model used in ChatGPT, introducing GPT 5.5 Instant with improvements in accuracy, personalisation and response brevity.
The updated model is being rolled out to all ChatGPT users and will replace GPT 5.3 Instant as the default option. GPT 5.3 Instant will remain available to paid users for three months through model configuration settings before being retired.
According to OpenAI, GPT 5.5 Instant is designed to provide shorter and more direct responses while improving factual accuracy and conversational tone. The internal evaluations showed the model produced 52.5% fewer hallucinated claims than GPT 5.3 Instant on prompts related to medicine, law and finance. It also reduced inaccurate claims by 37.3% in conversations previously flagged for factual errors.
The update also includes improvements in handling image uploads, STEM-related questions and deciding when to use web search to generate responses.
The model uses fewer words and lines in its replies compared with the previous version, while reducing unnecessary follow-up questions and excessive formatting.
The company is also expanding personalisation features in ChatGPT. GPT 5.5 Instant can use information from past chats, uploaded files and connected Gmail accounts, when enabled by users, to provide more tailored responses.
OpenAI also introduced a new ‘memory sources’ feature that allows users to see which saved memories or past conversations were used to personalise an answer. Users can delete or edit saved memories, remove chats from memory references or use temporary chats that do not store information.
Enhanced personalisation from past chats, files and connected Gmail accounts is rolling out to Plus and Pro users on the web, with mobile support planned later. The features will expand to Free, Go, Business and Enterprise users in the coming weeks, although availability may vary by region.
