After its videos generator from text or that can already use chatgpt as the default assistant on an Android mobile, OpenAi has announced its new most expensive reasoning model, the most expensive reasoning model, O1-PRO.
Since Deepseek appeared at the beginning of February, the responses of technological giants that develop larger language models are being diverse and OpenAI It is found in a complicated field, since since that date it has become Cancel O1, announce that it will take it to GPT-5 and various models as it happens now with O1-Pro.
And complicated because Sam Altman himself, CEO of OpenAI, declared that the idea of the company was to carry All its GPT-5 models and thus combine them in a new experience defined by the level of intelligence of the AI as the user paid and establish a free and without limits.
The new O1-PRO reasoning model, according to OpenAI, Use more computing power that the first O1 with the idea of offering more consistent answers. It is the big difference between the two and for the moment it will only be available for certain developers, those who have paid at least 5 dollars in OpenAi API services.
There is another very important point in this new model and it is Its pricesince Openai has raised it to make it quite prohibitive for some developments: $ 150 for one million tokens (750,000 words) introduced into the model and $ 600 per million generated by the model.
Which folds the price of GPT-4.5 per entrance and even increases it by 10 times compared to the O1 reasoning model. The high cost is due to the improved performance of O1-PRO, although OpenAi will have to convince developers to use it and that their effectiveness is really transferred to their developments.
A OpenAI spokesman has declared to Techcrunch: “O1-PRO in the API is an O1 version that uses greater computing power to think more depth and offer better responses to more complex problems. After receiving requests from our developer community, we are excited to incorporate it to give more effective answers.”
The free android
And according to the tests of this model of a few weeks ago, it is not very clear If it will be worth paying so muchsince it had a bit bad in the Puzzles Sudoku and with certain jokes of optical illusion in which it was not right at all. Where it has highlighted something has been in programming and mathematics against O1.
