Moonshot AI has released Kimi K2.6, an open-source native multimodal agentic model built on a Mixture-of-Experts architecture with 1T total parameters (32B activated). Key capabilities include long-horizon coding across Rust, Go, and Python; coding-driven UI/design generation; agent swarm orchestration scaling to 300 sub-agents and 4,000 coordinated steps; and proactive autonomous background execution. Benchmark results show competitive or leading performance against GPT-5.4, Claude Opus 4.6, and Gemini 3.1 Pro on SWE-Bench Verified (80.2%), SWE-Bench Pro (58.6%), LiveCodeBench v6 (89.6%), and agentic tasks. The model supports a 256K context window, native INT4 quantization, and deployment via vLLM, SGLang, or KTransformers. Both model weights and code are released under a Modified MIT License.