Renmin, Fudan Universities Propose CoworkOS for Persistent AI Coworkers
A joint team from Renmin University and Fudan University proposed CoworkOS, a system architecture for persistent, complex-task AI coworkers aligned with DeepSeek's Super Agent concept. It combines a three-layer collaboration architecture, shared application environments, and persistent work state with asynchronous execution. In tests across 12 real-world benchmarks and 6 delivery scenarios, CoworkOS performed comparably to DeepSeek Harness and Codex, and organized five iterative rounds of Qwen3.6-35B-A3B post-training, raising average scores from 33.19 to 42.95.
Coverage timeline
机器之心机器之心
近日，DeepSeek 为下一代超级智能体（Super Agent）提出了四个标准： 连续，高效，主动，持久，让智能体称为能够承担工作的 AI 同事。 与这一理念同频， 中国人民大学与复旦大学联合团队提出 CoworkOS， 将复杂任务协作、共享应用与持久工作状态纳入同一套系统。 论文标题：CoworkOS: Scaling Agent Coworkers for Real-world Automation 项目主页：https://coworkos-ai.github.io/ Demo Gallery：https://coworkos-ai.github.io/showcase.html 论文链接：https://github.com/RUC-NLPIR/CoworkOS/blob/main/CoworkOS_paper.pdf 代码地址：https://github.com/RUC-NLPIR/CoworkOS 下面的总览视频完整展示了 CoworkOS 如何工作以及对应的工作产物： 总览视频・CoworkOS 完整演示 理念同频之外，CoworkOS 进一步探索一种「超级 AI 同事」形态。这一探索沿着两个维度展开： 任务复杂度： 智能体需要通过多轮环境交互，来完成复杂办公交付工作。 时间跨度： 实现跨天、跨会话的持续工作，并在中断后接续推进。 CoworkOS 旨在让 AI 同事既能处理复杂任务，也能在更长的时间跨度内持续承担工作。 图 2：CoworkOS 总览：复杂办公交付与持续监控决策双挑战、三层协作架构与实验结果。 具体来说，CoworkOS 结合三项核心设计： 三层协作架构， 由主智能体制定策略与验收，协调智能体管理调度与状态，子智能体执行子任务； 共享应用环境， 支持人与智能体在 CAD、Blender 等应用中共同修订； 持久工作状态与异步执行， 通过工作区、记忆与后台任务支撑跨会话监控。 如图 2 所示，在 12 项真实世界基准和 6 大类交付场景中，团队将 CoworkOS 与 DeepSeek Harness、Codex 等系统进行对比，最终性能不分伯仲。此外，团队还以 模型自主后训练为 RSI 的探索场景， 让 CoworkOS 组织 Qwen3.6-35B-A3B 的五轮迭代，在八项基准上的平均分从 33.19 提升至 42.95。
Related stories
DeepSeek open-sources Super-agent architecture for autonomous bug-fixing agents
DeepSeek has publicly released the architectural blueprint for its Super-agent, a research prototype of a personal agent that maintains cross-session context to handle multi-task coordination. Internal tests show the agent can autonomously identify bugs and submit pull requests without being explicitly prompted, with a median task completion time of 12.4 minutes in simulated multi-task environments. DeepSeek positions this as a step toward reducing human attention bottlenecks in multi-task workflows.
3 reports · 3 sources
Fudan, Tencent Hunyuan, Tsinghua Release ExplorationBench to Test AI Rule Discovery
Fudan University, Tencent Hunyuan, and Tsinghua University released ExplorationBench, a benchmark measuring how AI systems discover rules in unfamiliar environments. It includes two altered 'alien worlds'—the programming language AlienCode and the logic proof system AlienLogic—where models must learn rules through trial and error. Results showed that without environment feedback, scores remained low (max 11.0%), while autonomous exploration led to high scores (up to 89.0%).
2 reports · 1 sources
Meta launches Muse Wonder and Muse Wonder Full AI models
Meta has released two new AI models, Muse Wonder and Muse Wonder Full, claiming state-of-the-art performance on personal and professional agentic tasks. The models introduce Active Memory and Active Focus features and are available to Muse Code subscribers, with Muse Wonder also accessible via the Meta Model API. Meta plans to eventually power its Muse AI agent with Wonder, starting with subscribers.
Apple and CMU propose SSR framework cutting multimodal search agent inference latency by over 90%
Researchers from Apple and Carnegie Mellon University propose SSR (Selection-based Structured Reasoning), a framework that replaces free-form token-by-token reasoning with selection from predefined reasoning options in multimodal search agents. In experiments across 7 multimodal search benchmarks, SSR reduces per-step inference latency by over 90% while maintaining task success rates comparable to free-form thinking.
News updates every hour. Subscribe in your RSS reader to get new stories automatically.
Subscribe