News","item":{"@type":"Thing","@id":"https://www.thestandard.com.hk/world-news"}},{"@type":"ListItem","position":5,"name":"Comment","item":{"@type":"Thing","@id":"https://www.thestandard.com.hk/comment"}}]}]17:T75b,

OpenAI will not release "o3" as a standalone artificial intelligence model, as the ChatGPT maker looks to simplify its products, CEO Sam Altman said on Wednesday.

The Microsoft-backed startup will release a GPT-5 model as a comprehensive AI system that will incorporate o3 along with other technologies, he said in a post on X.

OpenAI had unveiled o3 and o3 mini models in December 2024.

Its announcement comes at a time when companies in the U.S. are facing greater investor scrutiny over their massive spending on the technology, after Chinese startup DeepSeek unveiled a low-cost AI model last month.

"We want to do... a much better job simplifying our product offerings. We want AI to 'just work' for you; we realize how complicated our model and product offerings have gotten," Altman said, without sharing any timeline for the release of these products.

OpenAI did not immediately respond to a Reuters request for comment.

The generative AI pioneer aims to merge the o-series and GPT-series models in a bid to create AI systems that can utilize all available tools and handle variety of tasks.

OpenAI will also release a GPT-4.5 model, internally referred to as "Orion", as its last "non-chain-of-thought" model -- a reasoning process which generates direct answers without clear intermediate steps.

The model often struggles with complex reasoning tasks, particularly in domains such as physics and mathematics.

(Reuters)

OpenAI plans to simplify AI products in new road map for latest models, CEO Altman says