> ## Documentation Index
> Fetch the complete documentation index at: https://documentation.deepmask.io/llms.txt
> Use this file to discover all available pages before exploring further.

# Kimi K2 & K2.5 

> Kimi K2 and Kimi K2.5 on DeepMask. 1-trillion parameter MoE models with Agent Swarm Mode, multimodal reasoning, and 2M token context for complex agentic tasks.

MoonshotAI's Kimi K2 family introduces a new paradigm for agentic AI. Both models are built on a 1-trillion parameter Mixture-of-Experts architecture and natively support Agent Swarm Mode — coordinating up to 100 parallel sub-agents to decompose and execute complex tasks. Kimi K2 (DeepMask) is optimized for deep reasoning and long-context retrieval at 2M tokens, while Kimi K2.5 extends this with a focus on visual-to-code generation and full-stack prototyping from UI screenshots and video.

<Tabs>
  <Tab title="Kimi K2 (DeepMask)">
    ## About

    Kimi K2 (DeepMask) is MoonshotAI's massive Mixture-of-Experts breakthrough. It is the first model to natively support Agent Swarm Mode, allowing the main model to coordinate up to 100 specialized sub-agents working in parallel. Despite its 1-trillion total parameters, it activates only 32B per request — making it exceptionally efficient. It supports a 2M-token context window and uses MoonViT for native multimodal processing across text, images, and video.

    <Note>
      Kimi K2 (DeepMask) is available directly on DeepMask with a 2M-token context window and 256K-token support on standard hardware via Multi-Head Latent Attention.
    </Note>

    ## Key Capabilities

    <CardGroup cols={2}>
      <Card title="Agent Swarm Mode" icon="diagram-project">
        Decomposes a complex task into up to 100 parallel sub-tasks and executes them simultaneously.
      </Card>

      <Card title="Native Multimodal" icon="image">
        Processes text, images, and video with equal fluency through MoonshotAI's multimodal architecture.
      </Card>

      <Card title="2M Token Context" icon="arrows-left-right">
        Handles up to 2 million tokens, enabling analysis of massive documents and long-running agent sessions.
      </Card>

      <Card title="Stable Execution" icon="check-double">
        Maintains coherence across 300+ sequential tool calls without logic drift.
      </Card>
    </CardGroup>

    ## Use Cases

    * **Massive research synthesis** — Search hundreds of web sources simultaneously to compile comprehensive reports.
    * **Vision-to-code** — Upload a UI walkthrough video and have Kimi K2 rebuild the entire website structure.
    * **Batch data processing** — Analyze thousands of legal or medical records in a single swarm session.
    * **Multi-step agent workflows** — Run long-horizon tasks across tools, documents, and APIs without losing track of the goal.

    <Tip>
      Kimi K2 (DeepMask) is the strongest choice on DeepMask when you need to run complex, long-horizon tasks that span multiple tools and document types. Its 0.31s TTFT and 111 tokens/sec throughput make it fast enough for interactive use.
    </Tip>

    ## Specifications

    | Specification    | Value                          |
    | ---------------- | ------------------------------ |
    | Model Provider   | MoonshotAI                     |
    | Main Use Cases   | Visual Debugging, Agent Swarms |
    | Reasoning Effort | High                           |
    | GPQA Diamond     | 87.6%                          |
    | Max Context      | 2.0M Tokens                    |
    | Latency (TTFT)   | 0.31s                          |
    | Throughput       | 111 Tokens/sec                 |
  </Tab>

  <Tab title="Kimi K2.5">
    ## About

    Kimi K2.5 by MoonshotAI is a 1-trillion parameter MoE model built for the agentic era. It shifts from single-prompt interactions to a self-directed paradigm where it decomposes a project into parallel sub-tasks. Trained on 15 trillion mixed visual and text tokens, it is the top choice for visual-to-code generation, complex UI automation, and full-stack prototyping directly from screenshots or Figma links.

    <Note>
      Kimi K2.5 is a powerful open-source model. Its multimodal training on 15 trillion tokens makes it particularly strong for tasks that combine visual and code reasoning.
    </Note>

    ## Key Capabilities

    <CardGroup cols={2}>
      <Card title="Agent Swarm Orchestration" icon="diagram-project">
        Decomposes one goal into a coordinated team of sub-agents working in parallel across search, code, and terminal tools.
      </Card>

      <Card title="Coding with Vision" icon="code">
        Generates functional code directly from a UI screenshot or design link.
      </Card>

      <Card title="Visual Web Navigation" icon="globe">
        Browses the live web visually — clicking buttons and filling forms like a human user.
      </Card>

      <Card title="Multimodal Reasoning" icon="image">
        Integrates document analysis with real-time image search to solve data-heavy, cross-modal problems.
      </Card>
    </CardGroup>

    ## Use Cases

    * **Visual asset refinement** — Autonomously find, edit, and place assets into a web layout.
    * **Complex tool-augmented search** — Manage 100+ steps across search and terminal tools to verify scientific data.
    * **Full-stack prototyping** — Build functional web and mobile apps from zero to one, including backend logic.
    * **UI automation** — Navigate live interfaces and perform multi-step workflows like a human operator.

    <Tip>
      Use Kimi K2.5 when your task starts with a visual input — a screenshot, video walkthrough, or Figma design — and ends with running code or a deployed prototype. It is the strongest model on DeepMask for visual-to-code generation.
    </Tip>

    ## Specifications

    | Specification    | Value                                  |
    | ---------------- | -------------------------------------- |
    | Model Provider   | MoonshotAI                             |
    | Main Use Cases   | Visual Debugging, Coding, Agent Swarms |
    | Reasoning Effort | High                                   |
    | GPQA Diamond     | 87.6%                                  |
    | Max Context      | 262K Tokens                            |
    | Latency (TTFT)   | 0.52s                                  |
    | Throughput       | 138 Tokens/sec                         |
  </Tab>
</Tabs>