Occamy-1.0：面向协作的开源帕累托前沿35B智能
原标题：Occamy-1.0: Open Pareto-frontier 35B Intelligence for Co-work
AI 摘要
研究团队发布 Occamy-1.0，一个基于 Qwen3.6-35B-A3B 检查点继续训练得到的 35B 规模协作（co-work）智能体模型。该模型通过执行导向数据、可回放长程轨迹基础设施以及分阶段后训练（Marathon Expert 与 Sprint Expert 合并）来提升工具调用、状态跟踪与长任务连续性。在多个协作基准上，Occamy-1.0 在同等规模模型中表现领先，并在部分任务上可与更大前沿系统竞争，处于成本-性能帕累托前沿的低成本拐点。团队开源了模型权重和部分训练数据。
正文节选
Open Pareto-frontier 35B Intelligence for Co-work Abstract Co-work agents execute complex workflows that combine information gathering, tool use, coding, and file manipulation across many model invocations. Because cost and latency accumulate over the full episode, their practical value depends not only on peak capability but also on how efficiently that capability is delivered. Yet many steps in everyday work emphasize state tracking, coordination, recovery, and follow-through rather than front