- Kimi K3
- Kimi K2.7 Code
- Kimi K2.6 (evroc)
- Kimi K2 (DeepMask)
- Kimi K2.5
About
Kimi K3 is Moonshot AI’s newest model in DeepMask. It combines native vision, hybrid reasoning, and a 1M-token context window, so it can work across large document sets, screenshots, and long agent sessions in one conversation.Specifications
| Specification | Value |
|---|---|
| Model provider | Moonshot AI |
| Hosting | EU |
| Context window | 1M tokens |
| Max output | 262K tokens |
| Image input | Yes |
| Tool use | Yes |
| Reasoning | Hybrid reasoning |
About
Kimi K2.7 Code is a code-focused Kimi model. Use it for writing, reviewing, and refactoring code and for developer tasks that call tools. It works with text and documents, not images.Specifications
| Specification | Value |
|---|---|
| Model provider | Moonshot AI |
| Hosting | EU |
| Context window | 262K tokens |
| Max output | 262K tokens |
| Image input | No |
| Tool use | Yes |
| Reasoning | Standard |
About
Kimi K2.6 runs in Germany on evroc. It reads images and documents, supports reasoning and tool use, and is a strong choice when your work must stay on German infrastructure.Specifications
| Specification | Value |
|---|---|
| Model provider | Moonshot AI |
| Hosting | Germany (evroc) |
| Context window | 262K tokens |
| Max output | 262K tokens |
| Image input | Yes |
| Tool use | Yes |
| Reasoning | Built-in thinking |
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. In DeepMask, it is hosted in Germany, has a 262K-token context window, and works with text and documents.Kimi K2 (DeepMask) is hosted in Germany with a 262K-token context window. It doesn’t accept image input in DeepMask — use Kimi K3, K2.6 (evroc), or K2.5 for images.
Key Capabilities
Agent Swarm Mode
Decomposes a complex task into up to 100 parallel sub-tasks and executes them simultaneously.
Built-in Reasoning
Thinks through multi-step problems before answering, which helps on research and planning tasks.
262K Token Context
Handles up to 262K tokens, enough for long documents and extended agent sessions.
Stable Execution
Maintains coherence across 300+ sequential tool calls without logic drift.
Use Cases
- Massive research synthesis — Search hundreds of web sources simultaneously to compile comprehensive reports.
- 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.
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.
Specifications
| Specification | Value |
|---|---|
| Model Provider | MoonshotAI |
| Main Use Cases | Visual Debugging, Agent Swarms |
| Reasoning Effort | High |
| GPQA Diamond | 87.6% |
| Max Context | 262K Tokens |
| Latency (TTFT) | 0.31s |
| Throughput | 111 Tokens/sec |
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.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.
Key Capabilities
Agent Swarm Orchestration
Decomposes one goal into a coordinated team of sub-agents working in parallel across search, code, and terminal tools.
Coding with Vision
Generates functional code directly from a UI screenshot or design link.
Visual Web Navigation
Browses the live web visually — clicking buttons and filling forms like a human user.
Multimodal Reasoning
Integrates document analysis with real-time image search to solve data-heavy, cross-modal problems.
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.
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.
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 |