SemiAnalysis: Kimi K3 crushes Nvidia's strongest open-source model; U.S. committee model fails

```

Open competition among Chinese AI labs is bearing fruit.

Renowned semiconductor and AI research organization SemiAnalysis published an article on Wednesday pointing out that The latest Kimi K3 model from Moonshot AI has vastly outperformed Nvidia’s flagship open-source model Nemotron 3 Ultra in benchmark tests. SemiAnalysis directly named Nvidia CEO Jensen Huang, stating that the "Nemotron Committee" development model he led has proven not to be the right path for American open-source AI, and called on Nvidia to rethink its strategic direction.

SemiAnalysis believes that Chinese AI labs, through continuous iteration and rapid delivery in a free market, have formed a clear advantage in the open-source field. If the U.S. continues with the current committee-based collaboration model, it will continue to lag behind in this arena.

Committee Approach Creates "Groupthink"

SemiAnalysis attributes the failure of the Nemotron committee approach to its structural shortcomings. The institution stated that when Jensen Huang created the Nemotron Committee, this mechanism restricted the free flow of different technical routes and fostered groupthink—while the core value of open-source lies precisely in the ability for free experimentation, creating a fundamental contradiction between the two.

The problems don’t end there. SemiAnalysis revealed that Mistral made a mistake during Nemotron’s pre-training phase. Because Nvidia adopted a single committee structure, once a member company made an error, it dragged down the performance of the entire committee and triggered a chain of dissatisfaction among its members.

Even more concerning is the crisis of trust. SemiAnalysis stated that, according to its understanding, some alliance members were unwilling to share their best ideas, highest quality datasets, and evaluation schemes with the alliance, resulting in internal friction. The organization believes that committee-based training is "extremely clear" not to be the way forward.

Chinese Model Provides a Mirror

SemiAnalysis contrasted the competitive ecosystem of Chinese AI labs as a reference. The institution pointed out that while Chinese AI labs compete with each other, they also learn from proven successful experiences, resulting in different innovative outcomes such as KDA, MSA, and DSA.

In the view of SemiAnalysis, Kimi K3’s overwhelming lead over Nemotron 3 Ultra is a direct embodiment of the superiority of this competitive mechanism—Independent labs compete in testing different ideas, architectures, and technical routes, which is the only way to drive real technological progress.

SemiAnalysis Proposes Reform Roadmap

SemiAnalysis does not advocate Nvidia completely abandoning the committee model, but rather offers more specific improvement recommendations. The institution suggests that if Jensen Huang truly wishes to advance committee-based training, there should at least be three separate, mutually isolated independent committees to avoid groupthink.

These three alliances should independently innovate and compete in data, reinforcement learning, pre-training, and evaluation. SemiAnalysis believes such arrangements would better reflect American free market values and more closely mirror the actual operations of Chinese AI labs—iterating through competition, and differentiating real technological advances through iteration.

Risk Disclaimer and Exemption ClauseThe market involves risks, please invest with caution. This article does not constitute personal investment advice and does not take into account the special investment goals, financial circumstances, or needs of individual users. Users should consider whether any opinions, viewpoints or conclusions in this article are suitable for their particular situation. Investment based on this article is at your own risk. ```