Kimi K2.5
The premier "Agent Swarm" model designed for self-directed visual and text execution.

About the Model
Kimi K2.5 by Moonshot AI 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 can decompose a project into parallel sub-tasks. It is uniquely trained on 15 trillion mixed visual and text tokens, making it the top choice for visual-to-code generation and complex UI automation.
Model Key Capabilities
Agent Swarm Orchestration:
Decomposes one goal into a coordinated team of sub-agents working in parallel.
Coding with Vision:
Generates functional React/Tailwind code directly from a UI screenshot or Figma link.
Visual Web Navigation:
Browses the live web visually, clicking buttons and filling forms like a human user.
Multimodal Reasoning:
Seamlessly integrates PDF analysis with real-time image search to solve data-heavy problems.
Applications & Use Cases
Visual Asset Refinement:
Autonomously finding, editing, and placing assets into a web layout.
Complex Tool-Augmented Search:
Managing 100+ steps across search and terminal tools to verify scientific data.
Full-Stack Prototyping:
Building functional web and mobile apps "from zero to one" including the backend logic.
Recomended Models based on your needs
Model Specifications
General | |
|---|---|
Model Provider | MoonshotAI |
Main Use Cases |
|
Intelligence | |
Reasoning Effort | High |
GPQA Diamond | 87.6% |
Memory | |
Max Context | 262K Tokens |
Speed | |
Latency (TTFT) | 0.52s |
Throughput | 138 Tokens/sec |


