If you remember last November, November 2025, a lot of the changes happened there. New models from Anthropic and OpenAI has been released. All of us, at least me personally, I understood the change which has happened in AI space, not immediately, not at that moment, but more closer to February or even March 2026. Suddenly, lot of complex task became possible. You were able to basically just ask it for something, and the model was able to plan itself. So deep thinking was there, and the model then was just executing this plan.
And then you had an amazing result which was working from a first shot. What is happening now, I think, is the next level change, which we do not yet see or at least not all of us see, but I think it will be there, everyone will have an experience of its results. Maybe closer to a new year, maybe a bit after the new year let's say, in January or in February. I'm speaking about such models as Astra model from GPT, from OpenAI. And not only this model, I think the Fable 5.1, Fable 5 are at the same level, but I personally use Astra.
What I can tell you from my experience, why it's so much different and why I think it's the next stage. Here, we are talking about not just the model which can create plan, but a model which can execute on the goal which you define. And it normally do, at least for myself, it do a lot better than I am expecting the result to be. So I'm doing these videos. Normally, I record the video, and then I go back home. I use DaVinci Resolve Studio, then I cut the video, add some images.
It's quite a long process. It takes time. So what I asked Astra lately, I showed it my video in the beginning and my video in the end, and I asked how it can help me. It actually defined me not just the workflow and the process for editing this kind of videos with DaVinci Resolve MCP, which has just been released. It created even kind of recommendations how to improve the things and how to make the things faster. This is also, by the way, one of the reasons why I'm paying more for OpenAI models now than I was paying just two weeks ago.
I see that it saves me, at least for each video, at least one hour of my time. And if you quantify these hours of my time during the whole month, even if you just use it for the videos, it already makes sense to pay more than €20 or dollar for these models. So we are standing somewhere where the goal which you are giving and the result which is coming from the model is exceeding significantly, personally, my expectations. And I'm sure that it will be creating something which we are not yet used to in a very, very short amount of time.
