There have been recurring reports that OpenAI would release its next-generation AI model, GPT-5, this year. With the end of the year approaching, Sam Altman has finally spoken up, but unfortunately, we’ll have to wait until next year—GPT-5 is definitely not coming this year.
Image source:Wccftech
Sam Altman confirms no GPT-5 model will be released this year.
During a recent AMA on Reddit, OpenAI CEO Sam Altman revealed the company’s plans for this year and answered a wide range of user questions, including about GPT-5.
First, Sam Altman mentioned that OpenAI is currently focusing on launching o1 and its subsequent models:
我們目前正專注於推出 O1 及其後續模型。這些模型的複雜度都相當高，因此無法像預期中同時處理那麼多項目。（我們還面臨資源分配上的諸多限制，必須在許多優秀的構想中做出取捨。）至於 AVM 的視覺功能，我們目前尚無確切的發佈日期。
A laggymster also asked, “When will ChatGPT-5 or its equivalent be released, and what features will it have?” Sam Altman replied that better versions will come out later this year, but they won’t be calling it GPT-5.
Although Sam Altman didn’t explicitly say GPT-5 won’t be released this year, it’s basically confirmed from this Q&A that GPT-5 won’t come this year—at least we’ll have to wait until next year.
Additionally, the better version mentioned here is almost certainly the official release of o1, and there have recently been reports that OpenAI accidentally leaked it ahead of time.
Regarding the details, OpenAI also revealed a bit in the discussion thread—for example, compared to o1-preview, the official version will feature very noticeable improvements. Both o1 and GPT will continue to release new models in the future, though at some point they may start to merge. The o series models will also gain multimodal capabilities in the coming months.
The OpenAI CPO also revealed something I think is pretty important: autonomous task execution will be a major theme for ChatGPT in 2025.
For those who care about AI image generation, although OpenAI is making efforts, it is still unclear when a successor to Dalle 3 will arrive, which shows that AI image generation is not currently OpenAI’s main improvement task:
As for Sora, it is confirmed that its launch has been postponed. OpenAI has explained that the reason is that the model is still being refined, with a need to ensure the accuracy of safety, simulation safeguards, and other related features, as well as to expand computing power.
I guess the main issue is probably computing power—with the current hardware, it’s just not enough to accommodate the large number of users.
It actually makes sense that the GPT-5 model is being delayed, since the new o-series model was only just released as a preview version. It’s unlikely OpenAI would release another new model in such a short timeframe, and even if they did, the improvements would probably be minimal.
In any case, as can be seen from this discussion thread, next year’s ChatGPT is something to look forward to, especially in the area of automating tasks.
Source:Reddit
Source: KOCPC Chinese