Openai postpons the O3 AI model in favor of the “Unified” version of the next generation

Openai has effectively canceled the release O3Which was scheduled to be the next AI model for the company, in favor of what CEO Sam German calls a “simplified” product.

in Post in x on WednesdayAltman said that in the coming months, Openai will release a model called GPT-5 which “integrates a lot of (Openai) technology”, including O3, in the Chatbot platform that works in the IQ of artificial intelligence. Chatgpt And API. As a result of this road map decision, Openai is no longer planning to launch O3 as an independent model.

The company originally said in December it aims to launch O3 early this year. Just a few weeks ago, Kevin Whale, chief product official at Openai, He said in an interview O3 was on the right path to launch “February March”.

“We want to do a better job in sharing our intended road map, and a much better job simplifying our products offers,” Altman wrote in his position. “We want Amnesty International to” only “work for you; we are aware of the complexity of our model offers and products. We hate the models elected (in Chatgpt) as much as you do and want to return to Magic Unified Intelligence.”

Altman has also announced that Openai is planning to provide unlimited access to the GPT-5 in “Standard Intelligence Preparing”, according to “abuse sills”, as soon as the model is generally available. (Altman refused to provide more details about what this setting-these abuse

He added, “GPT-5 will include sound, fabric, research, deep search,” referring to a set of features launched by Openai in Chatgpt during the past few months. (A) The highest goal for us is to unify our (our) models by creating systems that can use all our tools, know when to think for a long time or not, and be generally useful for a wide range of tasks. “

Before GPT-5, Openai plans to launch GPT-4.5, which is “Orion”, which is “Orion”, in the next few weeks, according to Altman. Altman says that this will be the last “unsweetened model” of the company. Unlike other “thinking” models in O3 and Openai, unusual models tend to be less reliable in fields such as mathematics and physics.

Openai seems to be completely embracing the thinking model that can be said with the first thinking model, O1Late last year. Thinking models effectively verify themselves, which It helps them to avoid some of the pitfalls that usually rise. These fact examination of the facts bear some cumin-thinking models take a little longer, usually seconds longer, to reach solutions. But it tends to be more reliable and capable.

AI Lab Deepseek has recently acquired the world’s attention through the R1 model, which matches O1 to a number of criteria. Unlike O1, R1 is a “open” model under a mute license, which means that it can be downloaded and used as they see developers suitable.

In modern social media publications, altman I confess Deepseek has reduced Openai technology progress in artificial intelligence, and He said This Openai “pulls some versions” to compete better.

GPT-4.5, or OrionHe is said to have suffered a number of challenges related to performance and technical setbacks. Bloombergand InformationAnd Wall Street Magazine Independently stated that Orion showed less improvement in her predecessor, GPT-4OFrom GPT-4 did more than GPT-3.