Abstract created by Good Solutions AI
In abstract:
- OpenAI’s new GPT-5.5 Immediate mannequin turns into ChatGPT’s default, changing GPT-5.3 Immediate with considerably fewer gratuitous emojis in responses.
- PCWorld reviews the replace delivers 52.5% fewer hallucinations and 37.3% fewer inaccurate claims whereas offering extra concise solutions.
- Enhanced options embrace improved context integration from earlier chats, recordsdata, and Gmail, plus transparency displaying which reminiscence sources influenced responses.
One cause I took a break from ChatGPT a number of months in the past (I’m again now) was how sick to loss of life I acquired of its fixed emojis, particularly when it got here to lists. The mind emoji was my least favourite, together with the inexperienced checkmarks, the sharp fingers, and the yellow “hazard” indicators.
Effectively, I’ll imagine it after I see it, however with its newest “immediate” mannequin, OpenAI guarantees that we’ll be getting approach much less of these “gratuitous” emojis in ChatGPT’s responses.
Out there now for all ChatGPT customers—I haven’t seen it on my account but, so it should nonetheless be rolling out—GPT-5.5 Immediate is now the default ChatGPT mannequin, changing GPT 5.3 Immediate.
Amongst different qualities, OpenAI says GPT-5.5 Immediate is extra correct, serving up 52.5 % fewer hallucinations than GPT-5.3 Immediate on “high-stakes” prompts referring to medication, regulation, and finance, whereas slicing “inaccurate claims” by 37.3 % on “particularly difficult” chats. These are encouraging figures, however we’ll have to attend for unbiased benchmarks to judge its real-world efficiency.
The brand new mannequin also needs to supply “tighter and extra to-the-point” solutions “with out shedding substance,” OpenAI says. Maybe most significantly, GPT-5.5 Immediate “avoids issues that may make responses really feel cluttered, like gratuitous emojis.” Let’s hope so.
I’ve discovered that ChatGPT has been much less emoji-happy in current weeks, significantly when switching to the Professional mannequin (which, presently, is GPT-5.5). Nonetheless, the odd smiley or thumbs-up has nonetheless managed to look on occasion for the earlier “immediate” mannequin, versus Gemini and significantly Claude, that are virtually emoji-free.
Different options in GPT-5.5 Immediate embrace elevated effectiveness at pulling in context from earlier chats, recordsdata, and Gmail (offered you’ve enabled ChatGPT’s Gmail connector). The brand new mannequin will “intelligently” resolve whether or not so as to add personalization from different sources and is “quicker” at scouring outdated chats, OpenAI says.
Lastly, you’ll have the ability to evaluate which “reminiscence” sources had been utilized in a given chat irrespective of which ChatGPT mannequin you’re utilizing, which is nice for nixing a reminiscence supply that’s outdated.
I’ll have extra to say as soon as I’ve had an opportunity to provide GPT-5.5 Immediate a trial run, and hopefully we’ll have complete benchmarks from unbiased testers quickly. But when ChatGPT’s new “immediate” mannequin does handle to dial again the emojis, that’ll get a giant smiley from me—plus possibly a inexperienced checkmark or two.
