Golden Finance News, according to 21 Finance, on July 21, Huang Zhenxin, head of Kimi’s enterprise business, responded to market concerns. Huang Zhenxin clearly stated: “Open-source models, Chinese large models should not be tagged with low-price labels. We have built SOTA-level models, and we can also match reasonable commercial pricing.” In response to the speculation that “K3 is a distilled small model,” Huang Zhenxin clarified that the core reason for K3’s performance jump comes from underlying original architectural innovation, not from replicating distillation of existing models.
On the commercialization front, Huang Zhenxin revealed that Kimi does not currently offer private deployment services. Its business model benchmarks overseas mature top AI companies, focusing on honing core model foundation capabilities. In the company’s revenue structure, API business has the highest share. The company has not yet disclosed to the public the breakdown of revenue proportions for private deployment and industry solutions. At the same time, it also explicitly pointed out an industry parameter scale trend: the parameter scale of large models will continue to expand upward and will not stop at the current 2.8 trillion scale.
Regarding the debate between open-source and closed-source routes, Huang Zhenxin believes they are not opposing competitive relationships. Enterprise clients have differentiated needs: enterprises with needs for private deployment and fine-tuning tend to choose open-source, while those seeking stable hosted services choose closed-source. Open-source and closed-source themselves, as well as the quality of model tokens, are not bound together.
Since Kimi’s K2, it has continued to adhere to the open-source route, and this direction will not change in the future either. The core underlying technologies and papers are published, and the company maintains an open posture toward the industry. Regarding Musk’s remarks that Grok 4.5 may be able to surpass K3, Huang Zhenxin responded: “Let’s wait and see.”
When asked about the differences between Kimi’s open-source model and other domestic large models, Huang Zhenxin explained: the global open-source industry has standardized licensing agreements such as MIT. Kimi uses internationally recognized open-source licenses. The outside misunderstanding that Kimi cannot support private deployment stems from the fact that the current K3 has a huge parameter count and high deployment barriers, which is unrelated to the open-source licensing strategy itself.

View Original

This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.

Reward
like
Comment
Repost
Share

Comment

Add a comment

No comments

Trending Topics
View More
#
GUSDYieldRisesto3.8%
155.7K Popularity
#
EventContractsLive
73.14K Popularity
#
ETHBreaks1900
118.24M Popularity
#
SummerCreationCamp
1.19M Popularity
#
VIPExclusive4%APY
998.37K Popularity

Pinned

Sitemap

Kimi’s head responds to three major controversies: SOTA models should have their own pricing power

Trending Topics

GUSDYieldRisesto3.8%

EventContractsLive

ETHBreaks1900

SummerCreationCamp

VIPExclusive4%APY

Pinned