DownloadKimi K2.6 is an open-source native multimodal agentic model built for advanced autonomous execution, long-horizon coding, and large-scale task orchestration. It is designed to handle complex end-to-end software workflows across multiple languages and domains, including front-end development, DevOps, performance optimization, and coding-driven design. Beyond coding, it can transform prompts and visual inputs into production-ready interfaces and lightweight full-stack outputs with structured layouts, interactivity, and polished visual detail. One of its most distinctive capabilities is horizontal agent scaling, supporting up to 300 sub-agents and 4,000 coordinated steps in a single run, which enables parallel task decomposition and end-to-end completion of outputs such as documents, websites, and spreadsheets. Architecturally, it uses a 1T-parameter Mixture-of-Experts design with 32B activated parameters, a MoonViT vision encoder, and a 256K context window.
Features
- Native multimodal support for text, images, and video
- Long-horizon coding across multiple languages and domains
- Coding-driven design for production-ready interfaces
- Agent swarm orchestration with up to 300 sub-agents
- Supports up to 4,000 coordinated autonomous steps
- Preserve-thinking mode for stronger multi-turn reasoning
- 1T-parameter Mixture-of-Experts architecture
- Native INT4 quantization for more efficient deployment
Project Samples
