Although it is true that new developments are launched every few months in terms of artificial intelligence models, it seems that it is time to put on the brakes. For many, the reference continues to be ChatGPT, from OpenAI, since it is the AI that sets the trend and, above all, the one that millions of users use daily and there is news about it.
The current model, GPT-5, already makes it clear that things are pausing. A huge leap is expected, just as it already happened with GPT-4 and it certainly has not been the case. As often happens, once a new model is released, people begin to speculate when the next one will be released. Rumors about the next version, GPT-6, began to circulate… and the expectation has been growing.
Some media and analysts pointed to an imminent launch of GPT-6, even before the end of 2025. However, OpenAI has had to come out to stop the hype. Roon, the company’s engineer, did it first with a simple but resounding «No» in X, on October 18. Shortly after, an official spokesperson confirmed that there are no plans to launch GPT-6 in the short term.
OpenAI has made it clear that its focus is on improving GPT-5. The idea is not to create a completely new model, but to evolve the one that already exists, making it faster, more efficient and useful. Therefore, although there will be no GPT-6 this year (and probably not at the beginning of 2026), you can expect intermediate versions such as GPT-5.5, with improvements.
Instead of rushing with GPT-6, they prefer to refine what already works and add new features
Of course, it seems that 2025 and, above all, 2026, will be years of a certain pause. After the initial furor in the artificial intelligence race, expectations have collided with reality. AI experts are beginning to talk about what they have called the “valley of disillusionment”, a somewhat more cautious phase.
The AI bubble would continue to grow, but much more slowly. Companies would dedicate themselves to building better infrastructure and models that work at scale based on real use. That’s why, in the short term, forget about GPT-6.
Furthermore, it seems that OpenAI, with Sam Altman at the helm, have bigger problems to solve. The company accumulates million-dollar losses and the growth and advancement of ChatGPT only worsens this situation due to the investment.
Without going any further, they have already put more than a trillion dollars on the table to develop the necessary infrastructure, which includes 26 gigawatts of computing capacity backed by vendors such as Nvidia, AMD and Oracle.
The worst thing is that ChatGPT has already been ranked as the most profitable AI service. Since its debut, the ChatGPT app has been downloaded 690 million times, and has generated about 2 billion dollars. That sounds like a lot, but that’s only $2.91 per download..
Only about 5% of ChatGPT users pay for a monthly subscription that is around $20. Although OpenAI has the most used AI platform in the world, convincing the majority to pay is impossible seeing the good results of the free version, which also continues to improve.
The problem is that maintaining a service like this is not cheap: Operating costs alone are estimated to exceed $700,000 per day. And although OpenAI has doubled its revenue compared to the previous year, losses stand at $8 billion in the first half of 2025 alone. This leaves doubts in the air as to whether it is truly profitable.
«It seems crazy: We are currently losing money on OpenAI Pro (ChatGPT Pro) subscriptions! People are using it a lot more than we expected,» Altman commented in a tweet from January 2025.
With this as a basis, it is logical to think that it is time to put on the brakes and evaluate the current situation. A future GPT-6, following the logic, will require million-dollar investments by the company to make it a reality, but, most likely, history will repeat itself as with GPT-4 and GPT-5.
Don’t get carried away by it hype media, but consolidating what works and guaranteeing a positive return on all that investment seem to now be its two main objectives.
Get to know how we work in NoticiasVE.
Tags: Artificial intelligence, Software