我们推出了 Kimi K2.5，这是一个开源的多模态智能体模型，旨在推进通用智能体智能。K2.5 强调文本和视觉的联合优化，使两种模态相互增强。这包括一系列技术，如联合文本-视觉预训练、零视觉 SFT 和联合文本-视觉强化学习。在此多模态基础上，K2.5 引入了 Agent Swarm，一个自导向的并行智能体编排框架，能够动态地将复杂任务分解为异构子问题并并发执行。广泛评估表明，Kimi K2.5 在编码、视觉、推理和智能体任务等多个领域取得了最先进的结果。Agent Swarm 相比单智能体基线将延迟降低了多达 4.5 倍。我们发布了后训练的 Kimi K2.5 模型检查点，以促进智能体智能的未来研究和实际应用。
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.
核心贡献 · Key contributions
联合文本-视觉预训练采用早期融合与恒定比例，增强两种模态。 Joint text-vision pre-training with early fusion and constant ratio enhances both modalities.
零视觉微调仅用文本数据激活视觉推理与工具使用。 Zero-vision SFT activates visual reasoning and tool use using only text data.
视觉强化学习通过跨模态迁移提升文本性能。 Visual reinforcement learning improves textual performance via cross-modal transfer.