Home
How to Scale Large Models: Yang Zhilin's GTC Strategy on Token Efficiency and Agent Clusters

The ticket to the second half of the large model era is no longer about simply scaling compute, but a fundamental rethinking of the underlying architecture.
At the NVIDIA GTC 2026 conference on March 18, Moonshot AI founder Yang Zhilin delivered a highly anticipated keynote. This marked his first comprehensive public outline of the core technical roadmap behind the Kimi K2.5 model, providing fresh perspective on large model evolution in the "post-scaling" era.
Yang Zhilin stated that to break through current intelligence limits, a complete restructuring of key technologies like optimizers, attention mechanisms, and residual connections is essential. He framed Kimi's evolution across three synergistic dimensions:
Token Efficiency: Eliminating resource waste to pursue an even more extreme compute-to-performance ratio.
Long Context: Continuously deepening Kimi's long-context memory advantage to process information at a massive scale.
Agent Cluster: Intelligence is evolving from individual agents to dynamically generated "digital clusters."
In Yang Zhilin's view, scaling has now evolved into finding scale effects in efficiency, memory, and automated collaboration. Multiplying the gains from these three dimensions could unlock intelligence levels far beyond current capabilities.
According to earlier announcements, the Kimi K2.5 model launched in early January already demonstrates this "all-around" capability. As Moonshot AI's most powerful open-source model to date, it features a native multimodal architecture, achieves state-of-the-art (SOTA) performance in code and visual understanding, and supports flexible switching between "thinking" and "non-thinking" modes to precisely adapt to agent-based tasks.
As Moonshot AI's technological approach becomes clearer, the large model competition is shifting focus from "parameter count" to "intelligence density." With agent clusters emerging as a potential ultimate form of future intelligence, whether Kimi can achieve a breakthrough under Yang Zhilin's "three-dimensional multiplication" framework has become a key industry focus.
Related article
Talat’s AI meeting notes live on your device, not the cloud
Granola, the AI-powered notetaking app valued at $250 million, has gained traction among tech founders and venture capitalists. But one developer sees demand for a more private, fully local alternative available for a one-time fee with no subscriptio
New Roewe i6 Hits Market at 659,000 Yuan, Powered by Snapdragon 8155 and Doubao Large Model
SAIC Roewe today launched the new Roewe i6, a compact sedan that fully adopts the visual language of the Roewe D7. Its distinctive large upright grille and horizontal halo light bar stretch across the front, creating a strong sense of technology and
How to protect assets, buildings, and personal health?
In an unpredictable world, protection has become a strategic necessity—not just an option. Whether it's safeguarding finances, strengthening buildings, or focusing on personal health, long-term stability relies on proactive planning. True security is
Related Special Topic Recommendations
Comments (0)
0/500

The ticket to the second half of the large model era is no longer about simply scaling compute, but a fundamental rethinking of the underlying architecture.
At the NVIDIA GTC 2026 conference on March 18, Moonshot AI founder Yang Zhilin delivered a highly anticipated keynote. This marked his first comprehensive public outline of the core technical roadmap behind the
Yang Zhilin stated that to break through current intelligence limits, a complete restructuring of key technologies like optimizers, attention mechanisms, and residual connections is essential. He framed Kimi's evolution across three synergistic dimensions:
Token Efficiency: Eliminating resource waste to pursue an even more extreme compute-to-performance ratio.
Long Context: Continuously deepening Kimi's long-context memory advantage to process information at a massive scale.
Agent Cluster: Intelligence is evolving from individual agents to dynamically generated "digital clusters."
In Yang Zhilin's view, scaling has now evolved into finding scale effects in efficiency, memory, and automated collaboration. Multiplying the gains from these three dimensions could unlock intelligence levels far beyond current capabilities.
According to earlier announcements, the Kimi K2.5 model launched in early January already demonstrates this "all-around" capability. As Moonshot AI's most powerful open-source model to date, it features a native multimodal architecture, achieves state-of-the-art (SOTA) performance in code and visual understanding, and supports flexible switching between "thinking" and "non-thinking" modes to precisely adapt to agent-based tasks.
As Moonshot AI's technological approach becomes clearer, the large model competition is shifting focus from "parameter count" to "intelligence density." With agent clusters emerging as a potential ultimate form of future intelligence, whether
Talat’s AI meeting notes live on your device, not the cloud
Granola, the AI-powered notetaking app valued at $250 million, has gained traction among tech founders and venture capitalists. But one developer sees demand for a more private, fully local alternative available for a one-time fee with no subscriptio
New Roewe i6 Hits Market at 659,000 Yuan, Powered by Snapdragon 8155 and Doubao Large Model
SAIC Roewe today launched the new Roewe i6, a compact sedan that fully adopts the visual language of the Roewe D7. Its distinctive large upright grille and horizontal halo light bar stretch across the front, creating a strong sense of technology and
How to protect assets, buildings, and personal health?
In an unpredictable world, protection has become a strategic necessity—not just an option. Whether it's safeguarding finances, strengthening buildings, or focusing on personal health, long-term stability relies on proactive planning. True security is











