Key takeaways of zdnet
- Openai GPT-5 will be released “Soon”, which may be this week.
- It will automatically choose the best model for signals.
- It should help in giving high-quality and rapid reply.
Despite Openai launching its much awaited open -source model on Tuesday, people are already looking for the next big step of openi, with another big release rumble on the near horizon: GPT -5.
Also: Chatgpt can no longer ask you to break with your lover
While speculation about GPT-5 is not new, Easter eggs emphasize that this week can begin. Boris Power, Head of Applied Research in Openi, created one X post Read on Monday: “It is excited to see how the public receives GPT-5!”
Other companies are adding to rumors; For example, already Floeth Wish to, This GPT-5 is soon coming under its model picker. However, the most obvious evidence is the X post of CEO of OpenAII Sam Altman in which he shared, “Today was something big,” but a reference to the release of Open models, GPT-OSS-110B and GPT-OSS-20B, and “Then a large upgrade for GPT-5 after this week.
We have a lot of new items for you in the next few days!
Today something big-but small.
And then a large upgradation at the end of this week.– Sam Altman (@Sama) August 5, 2025
So, why the model has been making headlines for a possible release months? Whatever you should know here is updated with the latest information as we know.
Also: How to use chat: a starting guide for the most popular AI Chatboat
A quick eye on the timeline
When chats first popularity in November 2022, it was running at GPT-3.5. Openai then released the GPT-4 in March 2023, in which there was a big jump in the intellect at that time. Then, in March 2024, it introduced the GPT-4o, which improved its capabilities across the board and introduced multimodality.
Now, a year later, we are overdue for GPT-5, and it was not always a plan.
Also: Is the chatup plus really worth $ 20 when the free version provides so many premium features?
In February, Altman shared Roadmap On X for the model of the new company, and in the comments, he said that it would be a case “Week/month“Although in April, Ultman said the plans had changed, for the release of O3 and O 4-Mune and pushing the GPT-5 back for a few months. Altman claimed that the reason for the delay was due to challenges when all were trying to integrate all different elements in the model.
(Disclosure: ZDNET’s original company Ziff Davis filed a case of April 2025 against Openai, alleging that it violates Ziff Davis copyright training and operating its AI system.)
Recently, in mid -July, CEO of OpenaiI, Sam Altman said that GPT -5 was “soon being released” XThis was followed by a report Ruckus Citing insider sources, GPT-5 may be out in early August. With most of Openai’s model release, the report states that the release will be combined with mini and nano versions, which will also be made available to developers through an API.
So, what will really do GPT-5?
What do we know about GPT-5
According to Altman’s early roadmap, the GPT-5 O-Series models (think of O3 and O4 Reasoning models) and the GPT-Series models combine the arguments found in the model. The system should understand when the best use of various powers of both types of models, which will be a very big win for users, as it ensures that the user gets the best combination of speed, cost and quality.
While logic models are good in producing high-quality answers for complex problems, they take more time because they are thinking through the problem, as a result are more expensive, and efforts are not always necessary if the task does not demand it.
Also: This facility can make GPT-5 a true game changer (if Openai is correct)
On the other hand, if the chatgpt or user always oppose the GPT model by default, there are lost opportunities for the logic model to produce better reactions. Right now, users have the option to togue, but often people do not fully know which model will be compatible with their work.
Last week, Nick Turley, the head of the chatgipt, told me that the target of this release was that “the average person does not need to think which models to use.” In the context of a time -line, Turley also threw around the word “soon”.
In the same roadmap, Altman said that the Chatgpt-free users will get unlimited chat access to GPT-5 in Standard Intelligence Settings, while plus subscribers can reach a high level of intelligence level. Pro subscribers will be able to use it at the highest level. Models are slated to combine the best features of all chats, including voice, canvas, search, deep research, and more.
Also: How chat actually works (and why it is so game-changing)
To make expectations a bit, last week Altman reminds users that the model would be experimental and will include various research techniques, but it would not be the International Mathematics Olympiad (IMO) gold level, an award given to each other, still an anonymous model that is not expected to be released for months.
Want more stories about AI? Check out AI LeaderboardOur weekly newspapers.
You can put my latest stories and technical adventures on social media. Follow me on Twitter/X @sabrinaa_ortiz And on Instagram @sabrinaa.ortiz,
