Recently, it feels like AI models from various companies are being rolled out at an accelerating pace. OpenAI only just made the GPT-4.5 model available two months ago, and now foreign media reports have revealed that in addition to the previously mentioned o3 and o4-mini, OpenAI also has a model called GPT-4.1 on the way, which could very well be seen as early as next week. What’s more, this GPT-4.1 will come in multiple versions.
OpenAI could release models like GPT-4.1, o3 reasoning, and o4-mini as early as next week.
Previously, OpenAI had originally planned to release GPT-5 next, but last week OpenAI CEO Sam Altman said on social media that the plan had changed: they would first release o3 and o4-mini, and only then GPT-5. But now there might also be a GPT-4.1 as well.
change of plans: we are going to release o3 and o4-mini after all, probably in a couple of weeks, and then do GPT-5 in a few months.
there are a bunch of reasons for this, but the most exciting one is that we are going to be able to make GPT-5 much better than we originally…
— Sam Altman (@sama) April 4, 2025
According to a report from foreign media outlet The Verge, based on information obtained from insiders, OpenAI is currently preparing to launch multiple AI models. In addition to the already known o3 reasoning model and o4-mini, there is also one named GPT-4.1, a multimodal model expected to replace the current GPT-4o.
In addition, the GPT-4.1 model also appears to have smaller GPT-4.1 mini and nano versions, allowing users to switch between the model versions they need without having to use the largest GPT-4.1 model, thereby easing the current pressure on GPUs.
If this news is true, GPT-4.1 will replace GPT-4o, which also means that GPT-4.1 will support image generation at that point.
That said, Sam Altman previously mentioned that he understands how annoying it is to constantly switch between models, and that a unified model is planned for the future (tentatively to be realized with GPT-5). But now GPT-4.1 has been split into three versions, and on top of that there’s the o series and GPT-4o with scheduled tasks, so the number has only grown — it’s a bit awkward, really.
Earlier, AI engineer Tibor Blaho shared his findings from the new ChatGPT web version in the community, noting that the names o4-mini, o4-mini-high, and o3 appeared, seemingly hinting that these new models are about to launch. There are also reports that o3 and o4-mini will be unveiled next week, unless OpenAI adjusts its release plans.
New ChatGPT web app version deployed just now (after midnight San Francisco time) adds mentions of "o4-mini", "o4-mini-high" and "o3" pic.twitter.com/huz9XWQWGP
— Tibor Blaho (@btibor91) April 10, 2025
Although GPT-4.1 wasn’t mentioned, there’s a chance it’ll appear alongside them, since it’s part of the GPT series and doesn’t conflict with the o series.
Besides the new model, OpenAI also significantly improved earlier ChatGPT’s memory functionNow, ChatGPT can reference all your past conversations. This means that even if you didn’t ask ChatGPT to remember something, or it wasn’t automatically memorized before, as long as the conversation history is still there, ChatGPT can draw on it at any time and provide answers that better suit your habits and preferences based on the task at hand.
This will further enhance assistance in areas such as writing, seeking advice, and learning.
In the video, OpenAI demonstrates asking ChatGPT about previously discussed content regarding this weekend.
Starting today, memory in ChatGPT can now reference all of your past chats to provide more personalized responses, drawing on your preferences and interests to make it even more helpful for writing, getting advice, learning, and beyond. pic.twitter.com/s9BrWl94iY
— OpenAI (@OpenAI) April 10, 2025
Data source:The Verge
Source: KOCPC Chinese