OpenAI is rolling out a brand new ChatGPT mannequin Tuesday, specializing in making outputs extra helpful, private and correct. Based on OpenAI, GPT-5.5 Prompt will present clearer, extra succinct and fewer overly formatted solutions to on a regular basis queries. The brand new mannequin will restrict follow-up questions and “gratuitous emojis” to make responses much less cluttered.
The developer says GPT-5.3 Prompt hallucinates considerably lower than its predecessor on “high-stakes prompts” throughout matters like finance, regulation and drugs, and performs higher throughout picture reasoning, science and math.
Learn extra: OpenAI Made a Special ChatGPT for Your Doctor
Improved personalization and reminiscence management
OpenAI can also be leaning into improved personalization, saying GPT-5.5 Prompt can use previous chats, recordsdata and related Gmail accounts to reply prompts tailor-made to you, your workflow and tone.
The corporate can also be rolling out new reminiscence sources, supplying you with extra management, context and visibility into what data ChatGPT used to create the personalised responses.
“When a response is personalised, you’ll be able to see what context was used, akin to saved recollections or previous chats, and delete or right it if one thing is outdated or not related,” OpenAI says.
“Reminiscence sources aren’t proven to others should you select to share a chat. You stay in command of what’s in your reminiscence: you’ll be able to delete chats you not need to be cited, delete or change gadgets in saved recollections in settings or use momentary chats that do not use or replace your reminiscence.”
GPT-5.5 Prompt is out there Tuesday, changing GPT-5.3 Prompt because the default mannequin in ChatGPT and within the API as “chat-latest.” OpenAI mentioned GPT-5.3 Prompt will stay out there to paid customers for 3 extra months, after which will probably be retired.
Learn extra: ChatGPT 5.5 Is All About Math, Science and AI Research
