Pistis Technical Report

Abstract: We introduce the Pistis model family, comprising 27B- and 9B-parameter multimodal large language models built on Qwen3.6 and Qwen3.5, respectively, and developed through a general and scalable post-training framework.

摘要： 我们介绍了 Pistis 模型家族，其中包括分别基于 Qwen3.6 和 Qwen3.5 构建的 27B 和 9B 参数多模态大语言模型，这些模型是通过一个通用且可扩展的后训练框架开发的。

The framework first establishes a strong foundation through large-scale multimodal supervised fine-tuning (SFT). Building on this SFT foundation, we propose Interleaved Distillation and Reinforcement Learning (IDRL), a novel post-training paradigm that tightly integrates on-policy distillation and reinforcement learning within a single training loop.

该框架首先通过大规模多模态监督微调（SFT）建立了坚实的基础。在此 SFT 基础上，我们提出了交替蒸馏与强化学习（IDRL），这是一种新颖的后训练范式，它将策略内蒸馏（on-policy distillation）与强化学习紧密集成在同一个训练循环中。

By alternating between the two objectives, rather than optimizing either in isolation or combining them in a static joint loss, IDRL enables more effective knowledge transfer, greater optimization stability, and more precise credit assignment for long-horizon agentic trajectories, leading to stronger performance while mitigating common capability trade-offs.

通过在两个目标之间交替进行，而不是孤立地优化其中任何一个或将它们组合成静态的联合损失，IDRL 实现了更有效的知识迁移、更高的优化稳定性，并为长程智能体轨迹提供了更精确的信用分配，从而在减轻常见能力权衡的同时带来了更强的性能。

At both model scales, the framework produces two specialized variants: Pistis-Thinking, designed to strengthen deep multimodal reasoning, and Pistis-Agentic, which additionally incorporates agentic trajectory data to support long-horizon planning, iterative reasoning, and tool use. Pistis-Agentic is particularly strong in multimodal search. Both scales outperform their corresponding base models.

在两种模型规模下，该框架都产生了两个专门的变体：旨在加强深度多模态推理的 Pistis-Thinking，以及额外结合了智能体轨迹数据以支持长程规划、迭代推理和工具使用的 Pistis-Agentic。Pistis-Agentic 在多模态搜索方面表现尤为出色。两种规模的模型均优于其对应的基础模型。

Beyond model-parameter optimization, we further introduce Pistis-Auto-Harnessing (PAH), a system-level method that automatically improves the agent’s inference harness through iterative optimization. Experiments demonstrate that PAH enhances the model performance without updating the model parameters or increasing the interaction budget.

除了模型参数优化之外，我们还引入了 Pistis-Auto-Harnessing (PAH)，这是一种系统级方法，通过迭代优化自动改进智能体的推理工具（inference harness）。实验表明，PAH 在不更新模型参数或增加交互预算的情况下，提升了模型性能。