-
Advertisement
Artificial intelligence
TechTech Trends

Moonshot’s newest release narrows US-China AI model development gap: analysts

Kimi K2.5 raises questions about the potency of US controls on advanced chips to constrain China’s AI efforts, Brookings’ Kyle Chan says

Reading Time:2 minutes
Why you can trust SCMP
3
Chinese start-up Moonshot AI on Tuesday released its latest artificial intelligence model, Kimi K2.5. Photo: Shutterstock
Vincent Chow
Moonshot AI’s newest release, Kimi K2.5, has narrowed the gap between the US and China to the closest it has ever been in the development of artificial intelligence models, according to third-party evaluations.
That feat by the Beijing-based start-up, founded in March 2023, has raised questions about the efficacy of US policies – most notably, export controls on advanced semiconductors – to constrain China’s AI development efforts, said Kyle Chan, a fellow of the John L Thornton China Centre at the Washington-based think tank Brookings Institution.

“It seems that, for these Chinese start-ups, it was just a matter of getting access to capital,” Chan said. “I wouldn’t have guessed that Chinese AI companies would continue to keep pace with their US peers as recently as a month or two ago.”

Advertisement
Moonshot was valued at US$4.3 billion after raising US$500 million in its latest Series C funding round in December, according to venture capital tracker ITJuzi. Investors include IDG Capital, Tencent Holdings and Alibaba Group Holding, owner of the South China Morning Post.
Released on Tuesday, Kimi K2.5 scored just a handful of points behind the leading models of US heavyweights OpenAI, Anthropic and Google DeepMind in comprehensive performance evaluations by Artificial Analysis, the third-party benchmarking firm said on Wednesday.
Advertisement

Like other Chinese AI firms, Moonshot publicly released the weights of its new model, meaning that developers around the world with the requisite hardware could freely download and customise the powerful 595-gigabyte model without paying for expensive subscriptions.

The open-source nature of such a powerful model challenges the strategy of American AI companies that develop closed-source models, according to Chan.

Advertisement
Select Voice
Choose your listening speed
Get through articles 2x faster
1.25x
250 WPM
Slow
Average
Fast
1.25x