appConfigEntityId

bodyType

customContents

entityId

entityUuid

isVideo

articlePageQuery

article

articleSchemaContent

content

hooksArticlePageAdvertisementConfigArticle

hooksArticlePageFloatingButtonConfigArticle

appConfig

filter

json

liveArticle

Sam Phillips

Sam Altman, founder and CEO of ChatGPT owner OpenAI, said the release of its open-source artificial intelligence (AI) model would be delayed indefinitely, highlighting the challenges that the start-up faces in keeping pace with the open-source movement spearheaded by Chinese firms like DeepSeek.
While OpenAI’s closed-source AI models remain months ahead in their capabilities, Altman said on Saturday that its first open-source model in years may take more time to develop.
“We need time to run additional safety tests and review high-risk areas,” Altman wrote on his X account. “We are not yet sure how long it will take us.” The announcement echoes a similar one from last month, when he postponed a planned early summer release of the product.
The forthcoming model is distinct from previous releases like ChatGPT in that its source code and primary functions will be available for examination and modification by outside developers. It is an approach often adopted by companies seeking greater collaboration and more feedback on their products.

Users have been eagerly awaiting the release of this open-source model, particularly as Chinese AI companies, including DeepSeek and Alibaba Group Holding, are driving an open-source movement that is gaining traction around the world, including in the US. Alibaba owns the South China Morning Post.
Such an open-source strategy has facilitated the development of applications based on Chinese foundational models, posing a challenge to OpenAI. For example, DeepSeek R1, launched in January, offers capabilities similar to OpenAI’s paid models but is largely available for free, thanks in part to lower development costs.
“The curious thing about DeepSeek was really more that they were able to pull it off with the class of GPUs [graphics processing units] that they [used], which are not exactly the same as American GPUs,” said Shaw Walters, founder of US AI start-up ElizaOS.
“They were able to pull this off with what’s essentially 50 per cent less [GPU] performance.”
Walters suggested that DeepSeek’s surprise success may have motivated OpenAI to release its own open-source model.
“OpenAI has this killer model … [but] it’s not open source and everyone complains about it,” Walters said. “OpenAI has responded like, ‘OK guys, we’re going to release an open source model. We’ll do it. We hear you.’”
While Altman was uncertain about the timeline for the release, he assured those awaiting the update that his team was working diligently.
“This is new for us,” Altman wrote. “We want to get it right.”


China overtakes US as top workplace for elite AI researchers, study finds

China and the US are racing ahead on AI. Can they manage risks together?

Xi-Trump dinner puts US tech titans in spotlight – what does it mean for China ties?

From Suzhou to AI: Singapore, China eye new playbook for economic ties

Artificial intelligence (AI)

OpenAI delays its first open-source AI model in years challenging DeepSeek

OpenAI CEO Sam Altman. Photo: Reuters

Chinese companies such as DeepSeek have sparked a an open-source movement. Photo: AP

Meta’s AI lab is stacked with Chinese talent, drawing attention back home

Meta Platforms CEO Mark Zuckerberg makes a keynote speech during the Meta Connect annual event, at the company’s headquarters in Menlo Park, California, on September 25, 2024. Photo: Reuters

Chang Huiwen. Photo: LinkedIn

Lin Ji. Photo: Handout

Sun Pei. Photo: Handout

Zhao Shengjia. Photo: QQ.com

Bi Shuchao. Photo: Sina

Ren Hongyu. Photo: Stanford University

Yu Jiahui. Photo: QQ.com

Nvidia taps 2 young Chinese AI experts to strengthen research

Nvidia’s Jensen Huang. Photo: Reuters 

China catching up with US in algorithms despite chip gap, says former Microsoft AI head

Harry Shum Heung-yeung, former head of AI and research at Microsoft, speaks at an economic summit hosted by the HKU Business School on Friday. Photo: Handout



A DeepSeek poster seen at the Global Developer Conference in Shanghai in February. Photo: VCG via Getty Images

Chinese chipmaker Sophgo adapts compute card for DeepSeek in Beijing’s self-reliance push

A DeepSeek exhibit at the Hangzhou Future Science and Technology City exhibition hall in Hangzhou, Zhejiang province. Photo: CFOTO/Future Publishing via Getty Images

Tech

Big Tech

Tech Trends

Tech War

News

Business

Companies

China Future Tech

Founder and CEO Sam Altman says the release will be indefinitely delayed for additional safety tests


Founder and CEO Sam Altman says the release will be indefinitely delayed for additional safety tests.


OpenAI delays its first open-source AI model in years challenging DeepSeek

.css-1c6uqr6{color:inherit;font-weight:inherit;font-size:inherit;font-family:inherit;line-height:inherit;overflow-wrap:break-word;}Founder and CEO Sam Altman says the release will be indefinitely delayed for additional safety tests

Founder and CEO Sam Altman says the release will be indefinitely delayed for additional safety tests