ZDNET’s key takeaways
- OpenAI’s livestream at 10 AM PT/1 PM ET will likely launch GPT-5.
- GPT-5 will automatically select the best model for prompts, improving efficiency.
- That approach should help produce higher-quality answers more quickly.
OpenAI just launched its highly anticipated open-source models on Tuesday, but the company is already moving on to what will likely be its biggest product launch of the year: GPT-5.
Also: How ChatGPT actually works (and why it’s been so game-changing)
While speculation about GPT-5 isn’t new, there have been easter eggs nearly confirming that the release is taking place today. For starters, OpenAI cleverly wrote in an X post, “LIVE5TREAM THURSDAY 10AM PT” — replacing the “S” in livestream with the number five.
Then, in an even more obvious post, OpenAI shared a video on a video on X that said, “Dropping soon,” showing the “4” in GPT-4o moving on a dial to reveal the bottom of what looks like a five, as seen below.
Beyond OpenAI’s own teasing, several leaks from other companies confirm the timeline. First spotted by The Verge, Redditors noticed that GitHub accidentally announced OpenAI’s GPT-5 models in a now-deleted blog post. The post revealed that the family of models would include “gpt-5, gpt-5-mini, gpt-5-nano, and gpt-5-chat.” Earlier this week, AI platform Flowith teased that GPT-5 is coming soon under its model picker.
Also: How to use ChatGPT: A beginner’s guide to the most popular AI chatbot
So, why has the possible release of the model been making headlines for months? Here’s everything you need to know, updated with the latest information as we know it, and why you’ll want to tune into the live stream.
How to watch the live stream (and why you’ll want to)
You can watch the live stream by tuning in on the OpenAI site at 10AM PT/1PM ET, during which the company will share details of the models. OpenAI CEO Sam Altman said that it will be longer than normal, lasting around an hour, because there “is a lot to show.” However, if you can’t tune in, ZDNET will have all the news live here in real time.
our livestream tomorrow at 10 am PDT will be longer than usual, around an hour.
we have a lot to show and hope you can find the the time to watch!— Sam Altman (@sama) August 7, 2025
If you are debating whether it is worth watching, here’s a breakdown of what exactly GPT-5 will do.
What we know about GPT-5
When ChatGPT first rose to popularity in November 2022, it was running on GPT-3.5. OpenAI then released GPT-4 in March 2023, which at the time marked a big leap in intelligence. Then, in March 2024, it introduced GPT-4o, improving its capabilities nearly across the board, and introducing multimodality. Now, just over a year later, we are overdue for GPT-5.
Also: This one feature could make GPT-5 a true game changer (if OpenAI gets it right)
According to Altman’s initial roadmap, GPT-5 combines the reasoning capabilities found in the o-series models (think o3 and o4 reasoning models) and GPT-series models. The system should understand when to best use the different strengths of both types of models, which would be a huge win for users, because it ensures that the user gets the best combination of speed, cost, and quality.
While reasoning models are good at producing high-quality answers for complex problems, they take longer because they are thinking through the problem and are more expensive as a result. That effort isn’t always necessary if the task doesn’t demand it.
On the other hand, if ChatGPT or the user always opts for the GPT models by default, that’s a lost opportunity for the reasoning models to produce better responses. Right now, users have the option to toggle, but oftentimes, people don’t fully know which model would be better suited to their task.
Last week, Nick Turley, head of ChatGPT, told me the goal of this release was “that the average person does not need to think about which model to use.”
Also: Google’s Jules AI coding tool exits beta with serious upgrades – and more free tasks
In that same roadmap, Altman said that ChatGPT free users will get unlimited chat access to GPT-5 at the standard intelligence setting, while Plus subscribers could access a higher level of intelligence. Pro subscribers would be able to use it at the highest level. The models are slated to combine all of ChatGPT’s best features, including Voice, Canvas, Search, Deep Research, and more.
To temper expectations a bit, last week Altman reminded users that the model will be experimental and incorporate different research techniques, but it won’t be at International-Math-Olympiad-gold level, an award given to another, still-anonymous model that isn’t expected to be released for months.
Want more stories about AI? Check out AI Leaderboard, our weekly newsletter.
You can keep up with my latest stories and tech adventures on social media. Follow me on Twitter/X at @sabrinaa_ortiz and on Instagram at @sabrinaa.ortiz.
Leave a Reply