Kimi K2.5: Visual Agentic Intelligence

TL;DR

Kimi K2.5 is an open-source multimodal agentic model that integrates text and vision, featuring a novel agent orchestration framework, achieving state-of-the-art results and reduced latency in complex tasks.

Contribution

The paper introduces Kimi K2.5, a multimodal model with joint text-vision training and a new agent swarm framework for dynamic task decomposition and execution.

Findings

Achieves state-of-the-art performance in coding, vision, reasoning, and agentic tasks.

Reduces latency by up to 4.5 times compared to single-agent systems.

Provides open-source checkpoint for future research.

Abstract

We introduce Kimi K2.5, an open-source multimodal agentic model designed to advance general agentic intelligence. K2.5 emphasizes the joint optimization of text and vision so that two modalities enhance each other. This includes a series of techniques such as joint text-vision pre-training, zero-vision SFT, and joint text-vision reinforcement learning. Building on this multimodal foundation, K2.5 introduces Agent Swarm, a self-directed parallel agent orchestration framework that dynamically decomposes complex tasks into heterogeneous sub-problems and executes them concurrently. Extensive evaluations show that Kimi K2.5 achieves state-of-the-art results across various domains including coding, vision, reasoning, and agentic tasks. Agent Swarm also reduces latency by up to $4.5\times$ over single-agent baselines. We release the post-trained Kimi K2.5 model checkpoint to facilitate future research and real-world applications of agentic intelligence.