OpenAI says the new version of ChatGPT shows a bit more

The latest version of OpenAI’s ChatGPT is not capable of generating a lot of content. The hallucinations are has been an ongoing problem for AI models, but OpenAI says its new version of GPT-5.5 Instant has “significant improvements in real-time across the board.”

The company says that, based on “internal analysis,” GPT-5.5 Instant generated “52.5% fewer claims” than its Instant version of GPT-5.3 “on key issues affecting areas such as medicine, law, and finance.” GPT-5.5 Instant also reduced false positives by 37.3% for critical user-reported errors.

OpenAI also says that GPT-5.5 Instant is “more capable for everyday tasks,” such as analyzing uploaded images and knowing when to turn to the Internet for a response. GPT-5.5 Instant has “strong and extensive” solutions and avoids using “free emojis.”

With GPT-5.5 Instant, ChatGPT is now “more efficient” by taking things like past chats and your Gmail to provide you with personalized responses as well. (This is part of Google best selling Gemini Also.) And for all ChatGPT models, a new “memories” feature will allow chatbots to display the information that was used to inform a person’s responses, and you can delete or edit the information if you want.

OpenAI will begin rolling out GPT-5.5 Instant on Tuesday to “all ChatGPT users,” though GPT-5.3 Instant will be optional for three months until “retirement.” (In the past, users He is saddened by the death of former promotersso this gives people time to change.)

Enhanced personalization will begin for Plus and Pro users on the web and is “coming soon” to the mobile apps. OpenAI has “plans” to bring it “soon” to Free, Go, Business, and Enterprise users. The memories are running on the minds of ChatGPT consumers now online “and soon on the phone.”