AllSpark 发布开源搜索智能体 Iris-mini 与 Iris-pro
原标题：Iris-mini and Iris-pro are the strongest open-weight search agents in their class
AI 摘要
中国实验室 AllSpark 发布开源搜索智能体 Iris-mini（350 亿参数）和 Iris-pro（3970 亿参数），基于 Qwen 系列模型，支持 256K 上下文，并公开完整训练配方。训练数据从网页链接结构反向构造多步问题，经两阶段过滤和强化学习（SFT-RL climbing）优化。团队称两者在各自规模的开源搜索智能体中表现最强，并指出运行时上下文管理对基准成绩的影响常大于模型差异，且搜索训练还意外提升了通用工具使用和办公任务表现。
正文节选
Iris-mini and Iris-pro are the strongest open-weight search agents in their class The AllSpark team has released Iris-mini and Iris-pro, two open-source search agents along with a full training recipe. According to the paper, the training data and models also improved performance on tasks they were never trained for, including general tool use and office work. Search agents built on language models research the web on their own. They need to understand the question, decide what to search for, in