Well-designed from the start
Kimi K2.6 transforms simple prompts into Awwwards-level front-end interfaces. From high-impact hero sections, interactive elements, to scroll-triggered animations, every detail feels crafted rather than generated.
Kimi K2.6 is an open-source model featuring SOTA coding, long-horizon execution, and agent swarm capabilities, built to put them to work for you. Build complete products, run complex workflows through autonomous, coordinated agents, and create reusable skills that scale with you. One person can now achieve what once required a team.
Kimi K2.6 transforms simple prompts into Awwwards-level front-end interfaces. From high-impact hero sections, interactive elements, to scroll-triggered animations, every detail feels crafted rather than generated.
Go beyond the surface. Kimi K2.6 makes it easier to create user authentication, interactions, and database operations for lightweight use cases, so that you can build a complete, working website, all from a single prompt.
Handle complex, multi-step tasks with greater reliability. Kimi K2.6 stays aligned with your intent, makes fewer unnecessary changes, and corrects mistakes as it works, so you spend less time guiding and retrying.
Kimi K2.6 brings a major upgrade to the Agent Swarm experience. Agents now coordinate more effectively in parallel, combining strengths like broad search, deep research, large-scale analysis, long-form writing, and multi-format content generation. This improved coordination allows the swarm to complete deliverables across websites, documents, slides, and spreadsheets in a single run, with results that are more consistent, polished, and ready to use.
High-quality documents can now become reusable skills with Kimi K2.6, which captures how great work is structured and written. Apply these skills across future tasks to produce consistent, high-quality results without starting from scratch. Instead of repeating effort, your best work becomes something you can reuse, refine, and scale over time. When combined with Agent Swarm, these skills help produce more structured, consistent, and high-quality outputs across complex tasks.
Turn a single prompt and high-quality source material into a structured, multi-section astrophysics research project, where agent swarm collaborates to analyze, synthesize, and generate rich data-driven insights.
Create a Shakespeare calendar, with agent swarm coordinating content, visuals, and structure into a consistent, multi-format creative project.
Produce editorial-style visual projects with flexible layouts, rich imagery, and coordinated outputs across formats, powered by Agent Swarm and Document to Skills.
Develop multi-phase market strategies with structured outputs across PDF, PPT, and Excel, where agent swarm coordinates planning and Document to skills captures reusable strategic frameworks.
Turn a single prompt and high-quality source material into a structured, multi-section astrophysics research project, where agent swarm collaborates to analyze, synthesize, and generate rich data-driven insights.
Experience a new way to work with multiple agents as one coordinated team. Instead of managing each agent yourself, you can bring agents with different tools, contexts, and even models into a shared space, where they communicate, collaborate, and move tasks forward together. A coordinator helps organize the work by assigning tasks, managing dependencies, and keeping everything on track. You set the goal. The group handles the rest.
Kimi K2.6 now delivers stronger presentation generation, turning simple prompts or multi-format inputs into polished, professional slides. It can incorporate high-quality sources directly into your slides, improving both clarity and credibility. Upload screenshots or templates to create structured slides, with editable SmartArt like timelines, flowcharts, and funnels. The result is a fully editable deck ready to present.
Kimi K2.6 is our latest open-source model, featuring state-of-the-art coding, long-horizon execution, and agent swarm capabilities—the foundation behind everything you've seen.
Yes, Kimi K2.6 is free to use. If you'd like to access more features or enhance your workflow with advanced tools, paid plans are also available.
Go to kimi.com to get started. Using Document to skills is simple: Step 1: Click the “+” button, then go to Skills → Document to skills. Step 2: Upload your document and describe what you want Kimi to extract, replicate, or turn into a reusable skill. Step 3: Kimi analyzes the document and starts generating outputs based on your instructions. Step 4: Once complete, your results are ready to use and can be applied to future tasks.
Kimi K2.6 builds on previous models with stronger performance in coding, long-horizon execution, and agent swarm workflows. It handles complex, multi-step tasks with improved instruction following, higher reliability, and more stable long-running performance. Agents now coordinate more effectively in parallel, completing complex end-to-end deliverables across documents, websites, slides, and spreadsheets. It brings improved full-stack capabilities and new ways to turn high-quality documents into reusable skills for future tasks.
Claw Groups lets you work with multiple agents in a shared workspace. You can bring agents with different tools, contexts, and models together, while a coordinator assigns tasks, manages dependencies, and keeps the workflow on track.
Kimi K2.6 is available via the Kimi website, the Kimi App, the Kimi API, and Kimi Code.
Kimi K2.6 is an open-source model developed by Moonshot AI, featuring state-of-the-art coding, long-horizon execution, and agent swarm capabilities. It provides the core intelligence for a suite of professional tools, enabling you to build Awwwards-level interfaces, manage complex workflows with agent swarm, work with coordinated agents in Claw Groups, and deliver polished presentations with Kimi Slides.
Kimi K2.6 was developed by Moonshot AI, a company focused on building advanced AI models.
Yes, Kimi K2.6 is an open-source model. Its weights and code are publicly available on Hugging Face and GitHub, allowing developers and researchers to explore, integrate, and experiment with it.