What you need to know
- OpenAI just launched a new AI model dubbed GPT-4o with reasoning capabilities.
- A report suggests the ChatGPT maker might be getting ready to ship “GPT-5” later this year during the summer.
- Per a slide shown during ChatGPT voice demos at the VivaTech conference in Paris, OpenAI might be moving away from the traditional names and function across its AI models, consistent across GPT-1 through GPT-4, suggesting a new moniker.
OpenAI recently held its Spring Update event. While the company didn’t unveil the rumored smarter than the “mildly embarrassing GPT-4” GPT-5 or the long anticipated AI-powered search engine to compete with Bing and Google, it shipped its new ‘magical’ flagship model GPT-40 with reasoning capabilities across audio, vision, and text in real-time, making ChatGPT’s user experience more intuitive.
Sam Altman previously indicated that the nature of AI and software development dictates progression in performance capabilities when launching new models, as is the case with the chronological launch of GPT-1 to GPT-4. Each model is superior in every form when compared to its predecessor.
A common and almost obvious thought among most people is that OpenAI will launch GPT-5 as GPT-‘s successor. But as it turns out, this might not be the case. While showcasing ChatGPT’s overhauled voice experience at the VivaTech conference in Paris, OpenAI Head of Developer Experience Romain Huet presented a slide highlighting the growth of AI models over the next few years.
The slide showed the chronological launch of AI models from the GPT-3-era through to GPT-4. While the logical step representing a future iteration for the model would be “GPT-5,” OpenAI presented it as GPT-Next (via Tom’s Guide). Huet further disclosed that the company would ship multiple sizes of OpenAI models progressively over the years.
In my opinion, this might just be a placeholder while the ChatGPT maker settles on a more concrete moniker for the AI model. It almost reminds me of the Xbox ecosystem, jumping from Xbox 360 to Xbox One. Similar to what Microsoft is doing with its new Surface devices.
In a recent interview with the Director and GM of Redpoint, Logan Bartlett, Altman indicated GPT-5 may function like a ‘virtual brain’. Interestingly, he also revealed that model will ship under a “very special” name, further stating that its functionality will be very different from today’s models.
Everything we know about GPT-5 so far
A few days leading to OpenAI’s Spring Update event, Sam Altman cleared the air and revealed the company wasn’t launching GPT-5 or a new search engine. However, he indicated the team had been hard at work and stated that he was excited about the launch of new products, which felt “like magic” to him. While Altman indicated the company wasn’t going to launch GPT-5 at the event, he didn’t categorically indicate that it was out of the picture for the foreseeable future either.
OpenAI will eventually launch a new AI model with better capabilities, especially after a top OpenAI executive indicated that today’s ChatGPT will be laughably bad within the next 12 months. This lines up with Altman’s previous comments on the subject, where he indicated with a high level of scientific certainty that GPT-5 will be smarter than GPT-4, which he already admitted “kind of sucks.”
GPT-4 is the dumbest model any of you will ever have to use again by a lot. It’s important to ship early and often and we believe in iterative deployment.
OpenAI CEO, Sam Altman
Altman has also revealed that the AI revolution won’t require new devices, potentially indicating that the tech will be cloud-based. However, he claims, “You’ll be happy to have a new device.”
OpenAI has mastered the art of keeping things under wraps, but it hasn’t given a timeline for when we should expect these new AI models. A report suggests OpenAI is gearing up to ship a new model later this year during the summer that will reportedly make ChatGPT “really good, like materially better.”
