Kimi K2.7 Code:Built for Coding Agents
Kimi K2.7 Code is MoonshotAI's most capable coding model to date, tuned for long-horizon software engineering, tool calling, and agent workflows.
Built on the trillion-parameter K2 MoE backbone with 32B active parameters, Kimi K2.7 Code supports a 256K context window, text/image/video input, automatic context caching, and always-on thinking for multi-step engineering work.
Kimi K2.7 Code Quick Experience
Try the coding-focused AI assistant
Kimi K2.7 Code is live. I can work with long code context, keep thinking enabled for multi-step tasks, and help with debugging, refactoring, tests, and agent workflows. What are you building?
Kimi K2.7 Code Improves Long-Horizon Coding
The official model card reports K2.7 Code at 62.0 on Kimi Code Bench v2, 53.6 on Program Bench, and 35.1 on MLS Bench Lite, improving over K2.6 on all three.

Agentic Coding
Plan, edit, test, and iterate across real projects
Stronger Coding Benchmarks
Improved Kimi Code Bench v2, Program Bench, and MLS Bench Lite scores
MoE Architecture
1T total parameters with 32B active per token
Key Features of Kimi K2.7 Code
A coding-first K2 release for long-context software engineering, multimodal requirements, and agentic developer workflows.
What is Kimi K2.7 Code?
Kimi K2.7 Code is MoonshotAI's coding-focused K2 model for code generation, codebase understanding, tool use, and agentic software engineering. It is positioned as Kimi's strongest coding model to date.
About Kimi K2.7 Code
Kimi K2.7 Code is a coding-specialized model built on the Kimi K2.6 foundation. It targets real software engineering workflows: codebase understanding, long-context instruction following, multimodal requirements, tool calls, and agentic task execution.
The release keeps the K2 family MoE backbone: 1T total parameters, 32B active parameters, 384 experts, 61 layers, MLA attention, SwiGLU activation, and a 256K token context window. The Hugging Face model card also lists MoonViT as the vision encoder for image and video workflows.
For developers, the practical entry points are Kimi Code, the Kimi API, OpenRouter, and compatible coding tools. K2.7 Code is most relevant when the task involves debugging, refactoring, test generation, multi-file edits, project analysis, or long-running coding agents.
K2.7 Code Technical Specs
- • 256K context window for codebases and logs
- • Always-on thinking; disabled thinking is not supported
- • Text, image, and video input
- • ToolCalls, JSON Mode, Partial Mode, context caching
- • 1T / 32B active MoE architecture, modified MIT weights
K2.7 Code Use Cases
- • Long-context codebase understanding
- • Debugging, refactoring, and test generation
- • Screenshot-to-UI and video-based issue analysis
- • Claude Code, Cline, RooCode, and Kimi Code workflows
- • Custom AI coding assistants via API
Kimi K2.7 Code Signals
Benchmark and workflow highlights for developers evaluating K2.7 Code in real coding-agent tasks.
"The multimodal path matters: screenshots, design notes, and code can sit in one task, which makes K2.7 Code a better fit for UI implementation and review loops."
"The 256K context window is the practical feature for codebase work. It gives the model room to keep API contracts, logs, tests, and implementation files together."
"Always-on thinking is a sensible default for coding agents. Debugging and refactoring need planning, failed-attempt recovery, and tool feedback more than a fast first token."
"The official MLS Bench Lite lift is interesting because many coding models struggle when code, experiments, and ML reasoning are mixed in the same workflow."
"Kimi Code plus K2.7 Code is the workflow to watch: ask it to inspect the repo, make a small change, run the test loop, and explain the tradeoff."
"The OpenAI-compatible API and OpenRouter model ID make it straightforward to evaluate K2.7 Code without redesigning an existing model-routing layer."
"Kimi K2.7 Code is best understood as a coding-agent model: long context, always-on thinking, multimodal inputs, and tool-oriented execution packaged for real developer workflows."
Kimi K2.7 Code FAQ
Answers to common questions about Kimi K2.7 Code, Kimi Code, context length, thinking mode, pricing, and access.
Need technical support?
Access documentation, model weights, and developer resources for Kimi K2.7 Code.
Kimi K2.7 Code resources: API Documentation • HuggingFace • GitHub