When I launched Openai GPT-5 Last week, the company said the model will simplify the ChatGPT experience. Openai expressed the hope that the GPT-5 will serve as a type of “one size that suits everyone” with a router that would automatically decide the best way to answer user questions. The company said that this unified approach will remove the need for users to move Openai Sam Altman, CEO of Openai Sam Altman, said he hates.
But it seems that the GPT-5 is not the Openai-Unified IQ model.
Al -Taman said in a After x Tuesday that the company introduced the “Auto”, “Fast” and “Fast” and “Thinking” settings for the GPT-5 that all Chatgpt users can choose from the models. It seems that the preparation of cars is like the GPT-5 model that announced Openai at the beginning; However, the company also gives users options to move in, allowing them to reach artificial intelligence models quickly and quickly directly.
Besides the new GPT-5 situations, Altman said that the paid users again can reach many old AI’s models-including the GPT-4O, GPT-4.1 and O3- that were neglected last week only. GPT-4O is now in virtual models, while other Amnesty International models can be added from ChatGPT settings.
“We are working on updating the GPT-5 character that should feel warmer than the current but not annoying person (for most users) like GPT-4O”, Altman wrote in the post on X. “However, one of the past few days learns to us that we really need to reach a world with more user allocating to the typical character.”
Chatgpt’s conferees seem to be as always as always as it was always, indicating that the GPT-5 model router has not satisfied the world as the company was contemplating. The expectations for GPT-5 were high, in the hope that Openaii would push the boundaries of artificial intelligence models such as GPT-4 launch. However, the GPT-5 width was more severe than expected.
I neglect GPT-4O and other Amnesty International models in Chatgpt A. A violent reaction between users Those who grew associated with the responses and models of artificial intelligence in ways not expected by Openai. In the future, Altman says that the company will provide users a lot of prior notice if it is ever from GPT-4O.
TECHRUNCH event
San Francisco
|
27-29 October, 2025
It seems that the GPT-5 model router has been largely broken on the launch day. This caused some users to feel that the artificial intelligence model was not like the previous Openai models, and I was forced to address the problem in Ama session on Reddit. However, it seems that the GPT-5 router may still be unjust to all users.
“We will not always get everything in attempt No. 1, but I am very proud of the speed of the team’s repetition.” After x Tuesday.
The direction of claims to the correct artificial intelligence model is a difficult task that requires aligning the artificial intelligence model of the user’s preferences, as well as the specific question that it asks. After that, the router must take a decision on the artificial intelligence model to send its claim in only one second – in this way, if you go to a fast -responding AI model, the response can be fast.
On a wider scale, some people offer preferences to artificial intelligence models that exceed rapid or slow responses. Some users may admire the AI model, while others may estimate contradictory answers to another.
Human association with some artificial intelligence models is a relatively new concept that is not well understood. For example, hundreds of people in San Francisco recently A funeral was held For the AI model of anthropology, Claude 3.5 Sonata, when it was taken in a non -connection mode. In other cases, chat groups appear to be artificial intelligence Mention rabbit holes.
OpenAi seems to have more work to do it around the artificial intelligence models with individual user preferences.
https://techcrunch.com/wp-content/uploads/2024/10/GettyImages-2174797696.jpg?resize=1200,800
Source link