Topics
Latest
AI
Amazon
Image Credits:Eugene Gologursky/The New York Times / Getty Images
Apps
Biotech & Health
Climate
Image Credits:Eugene Gologursky/The New York Times / Getty Images
Cloud Computing
Commerce
Crypto
Enterprise
EVs
Fintech
Fundraising
gadget
game
Government & Policy
Hardware
layoff
Media & Entertainment
Meta
Microsoft
Privacy
Robotics
Security
Social
Space
Startups
TikTok
transport
Venture
More from TechCrunch
Events
Startup Battlefield
StrictlyVC
Podcasts
TV
Partner Content
TechCrunch Brand Studio
Crunchboard
Contact Us
OpenAI has effectively canceled the sacking ofo3 , which was slated to be the company ’s next major AI model , in favor of what CEO Sam Altman is call a “ simplified ” product offering .
In apost on X on Wednesday , Altman say that in the coming month , OpenAI will release a model called GPT-5 that “ integrates a lot of [ OpenAI ’s ] technology , ” let in o3 , in its AI - power chatbot platformChatGPTand API . As a result of that roadmap decisiveness , OpenAI no longer plans to launch o3 as a stand - alone model .
The company originally said in December that it propose to issue o3 sometime early this year . Just a few weeks ago , Kevin Weil , OpenAI ’s principal product officer , said in an interviewthat o3 was on racetrack for a “ February - March ” launch .
“ We want to do a better job of partake our intended roadmap , and a much better caper simplifying our production offerings , ” Altman wrote in his C. W. Post . “ We want AI to ‘ just work out ’ for you ; we realize how complicated our role model and mathematical product offering have produce . We hate the model chooser [ in ChatGPT ] as much as you do and want to return to wizard incorporated word . ”
Altman also announced that OpenAI project to offer unlimited shoot the breeze access to GPT-5 at the “ standard intelligence background , ” subject to “ abuse thresholds , ” once the model is broadly available . ( Altman declined to furnish more detail on what this setting — and these revilement thresholds — entail . ) contributor to ChatGPT Plus will be able-bodied to run GPT-5 at a “ high level of intelligence , ” Altman said , while ChatGPT Pro reader will be able-bodied to fly the coop GPT-5 at an “ even higher level of intelligence . ”
“ [ GPT-5 ] will contain voice , canvass , search , deep enquiry , and more , ” he added , referring to a chain of mountains of features OpenAI has plunge in ChatGPT over the preceding few calendar month . “ [ A ] top destination for us is to unify [ our ] models by creating organisation that can use all our tools , have a go at it when to guess for a long time or not , and generally be useful for a very wide reach of task . ”
Before GPT-5 stray out , OpenAI plans to release GPT-4.5 , a model code - named “ Orion , ” in the next several weeks , according to Altman . Altman says this will be the company ’s last “ non - Sir Ernst Boris Chain - of - think model . ” Unlike o3 and OpenAI ’s other “ abstract thought ” models , non - chain of mountains - of - thought poser tend to be less reliable in domains like mathematics and physics .
Join us at TechCrunch Sessions: AI
Exhibit at TechCrunch Sessions: AI
It seems that OpenAI is fully embrace the logical thinking model drift it arguably kickstarted with its first abstract thought model , o1 , late last year . abstract thought model effectively fact - check themselves , whichhelps them to avoid some of the pitfalls that normally trip up example . This fact - check process incurs some latency — reasoning models take a small longer , usually seconds to minutes longer , to arrive at solutions . But they incline to be both more reliable and capable .
Taiwanese AI research lab DeepSeek capture the world ’s attention recently with its R1 model , which matched o1 on a number of bench mark . As opposed to o1 , R1 is an “ open ” example under a permissive permit , meaning it can be downloaded and used as developers see set .
In late social media post , Altmanadmittedthat DeepSeek has subside OpenAI ’s technical lead in AI , andsaidthat OpenAI would “ pull up some releases ” to best compete .
GPT-4.5 , or Orion , is said to have suffered a act of performance - link up challenges and technical setback . Bloomberg , The Information , andThe Wall Street Journalhave independently reported that Orion has shown less of an improvement over its precursor , GPT-4o , than GPT-4 did over GPT-3 .
