Kimi K2.5 is a powerful open-source model that delivers state-of-the-art coding and vision capabilities, along with a self-directed agent swarm paradigm. It builds on Kimi K2 with continued pretraining over approximately 15T mixed visual and text tokens, enabling it to excel in complex tasks. The model can self-direct an agent swarm with up to 100 sub-agents, executing parallel workflows across up to 1,500 tool calls, reducing execution time by up to 4.5x compared to single-agent setup.
Kimi K2.5 excels in coding with vision, allowing it to turn simple conversations into complete front-end interfaces with interactive layouts and rich animations. It also improves image/video-to-code generation and visual debugging, making it easier for users to express intent visually. The model is available via Kimi.com, the Kimi App, API, and Kimi Code, with four modes: K2.5 Instant, K2.5 Thinking, K2.5 Agent, and K2.5 Agent Swarm (Beta), offering a range of options for users to interact with the model.
Kimi K2.5 brings agentic intelligence into real-world knowledge work, enabling it to handle high-density, large-scale office work end-to-end. It reasons over large inputs, coordinates multi-step tool use, and delivers expert-level outputs, including documents, spreadsheets, and slide decks, directly through conversation. The model supports advanced tasks, such as adding annotations and constructing financial models, and scales to long-form outputs, making it a powerful tool for real-world professional tasks, with tasks that once took hours or days now completing in minutes.
Get more likes & reach the top of search results by adding this button on your site!
Get top updates in AI to your inbox every weekend. It's free!