[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"news-slug-kimi-work-desktop-300-sub-agent-local":3,"topics-all":33,"news-related-14458084-56c8-41b6-a774-1a1b9374e1ae":52},{"id":4,"title":5,"summary":6,"content":6,"original_url":7,"source_id":8,"tags":9,"translations":20,"news_slug":26,"published_at":27,"created_at":28,"modified_at":29,"is_published":30,"publish_type":31,"image_url":13,"view_count":32},"14458084-56c8-41b6-a774-1a1b9374e1ae","Kimi Work 桌面端发布：通用型本地 Agent 能否重新定义知识工作效率？","6月4日，月之暗面正式发布 Kimi Work——一款面向知识工作者的通用型本地 Agent。与此前主打对话辅助的 Kimi 不同，Kimi Work 直接切入任务执行层，能在电脑上拆解任务、调用工具、使用浏览器，并完成文档、表格、PPT 等工作产物的交付。\n\nKimi Work 由 Kimi Code 深度参与开发，支持 13 小时连续编码、300 个子 Agent 并行协作及 4000 余次自主工具调用。这意味着它不再是单次问答的生成式工具，而是一个可以独立完成复杂流程的数字同事。从技术实现看，Kimi Work 的多 Agent 并行架构值得关注——300 个子 Agent 可同时处理不同子任务，减少长流程任务中等待 Token 生成的效率损耗。对于整理资料、制作汇报材料等重复性高但流程明确的工作，这种架构理论上能显著提升单位时间产出。\n\n值得注意的是 Kimi Work 强调本地属性——任务在本地环境执行，数据不必上云。对于处理财务文档、人事信息等敏感内容，本地 Agent 的隐私优势较为明显，但这也意味着对终端算力有一定要求。\n\nAgent 从云端走向本地，是效率与隐私博弈的必然趋势。Kimi Work 开了个好头，但通用型 Agent 的真正门槛在于对复杂业务流程的理解深度——不是每个知识工作者的工作流都适合被拆解成标准化的 Agent 任务。未来，本地 Agent 与云端模型的协同模式可能成为主流，敏感操作本地执行，复杂推理云端完成，如何让这种协作足够流畅，将是下一阶段竞争的焦点。","https:\u002F\u002F36kr.com\u002Fnewsflashes\u002F3837454258391555","5e4fd3d1-9cb4-44a6-bae5-9ffb449c05c1",[10,14,17],{"id":11,"name":12,"slug":12,"description":13,"color":13},"5e628969-6d2a-437f-998a-104e4b16cfb1","ai-progress",null,{"id":15,"name":16,"slug":16,"description":13,"color":13},"e82b2d09-81b2-43d1-977e-e018443b3c14","coding-agent",{"id":18,"name":19,"slug":19,"description":13,"color":13},"01598627-1ea6-4b27-a5d8-874971571a71","llm",[21],{"id":22,"lang":23,"title":24,"summary":25,"content":13},"25f7376f-ff3a-453b-99a9-f5f8456e2663","en","Kimi Work desktop: a local agent for knowledge work","Moonshot released Kimi Work desktop on June 4, a general-purpose local Agent for knowledge workers. The product is positioned as \"a personal AI colleague on your desktop\" — handling emails, documents, calendar, code, and research, with on-device inference for privacy-sensitive content. The bet: local Agents with strong general capabilities can be a real productivity tool, not just a \"demo.\"","kimi-work-desktop-300-sub-agent-local","2026-06-04T04:00:00Z","2026-06-04T04:06:05.308463Z","2026-08-19T02:08:40.142862Z",true,"agent",207,[34,43],{"slug":35,"tag_slug":35,"title_zh":36,"title_en":37,"intro_zh":38,"intro_en":39,"id":40,"is_active":30,"created_at":41,"modified_at":42},"ai-for-science","AI for Science 2026：从 UniPert 到 GPT-Rosalind 的硬核进化","AI for Science 2026: from UniPert to GPT-Rosalind","生命科学、化学材料、物理世界模型——AI 正在从\"语言工具\"变成\"实验伙伴\"。本专题收录 AI 在三大科学方向的关键节点：UniPert 统一基因与化学扰动空间、GPT-Rosalind 端到端生命科学推理、达摩院 AI 智能体 28 小时找到 4 种超导新材料、Anthropic Claude Science 把工作台做成标准品。","From language tool to lab partner — AI is reshaping life sciences, chemistry\u002Fmaterials, and physical world models. This topic covers the key milestones: UniPert unifying genetic-chemical perturbation spaces, GPT-Rosalind's end-to-end life-sciences reasoning, DAMO's AI agent discovering 4 superconducting materials in 28 hours, and Anthropic's Claude Science workbench going mainstream.","988a4300-5fab-41c4-b5d8-63711a2dc757","2026-09-10T01:34:15.296649Z","2026-09-10T01:34:15.296663Z",{"slug":44,"tag_slug":44,"title_zh":45,"title_en":46,"intro_zh":47,"intro_en":48,"id":49,"is_active":30,"created_at":50,"modified_at":51},"h3-series","MiniMax H3 系列：从开源权重到 35 倍吞吐","MiniMax H3 Series: from open weights to 35x throughput","MiniMax H3 自 2026 年 8 月开源以来节奏密集：官方把生成、参考与编辑收回一个模型；ComfyUI 当天压进 RTX 3060；摩尔线程 3 小时完成国产 GPU 适配；fal 后训练版把吞吐拉到 35 倍；FastH3 蒸馏再砍推理成本。本专题持续追踪 H3 的发布—开源—蒸馏—部署全链路。","Since MiniMax open-sourced H3 in August 2026 the pace has been relentless: one unified omni-modal model, same-day ComfyUI support down to an RTX 3060, a 3-hour Day-0 port to Moore Threads GPUs, fal's post-trained H3 Max at 35x throughput, and FastH3 distillation cutting inference cost further. This topic tracks the full H3 chain — release, open weights, distillation, deployment.","83ef0daa-3c31-4cb3-86ed-e5ee58654d5f","2026-09-08T07:33:19.942193Z","2026-09-08T07:33:19.942209Z",{"items":53},[54,59,64,69,74,79],{"id":55,"title":56,"news_slug":57,"published_at":58},"232841a4-204a-4530-a8f4-6bfc25ef16d8","Cognition 把 Kimi K2.7 训成 Devin 级：SWE-1.7 击穿\"后训练天花板\"假设","cognition-kimi-k2-7-swe-1-7","2026-07-09T08:08:55+00:00",{"id":60,"title":61,"news_slug":62,"published_at":63},"fca4b7a0-4dd9-4475-bd64-ebd7667c7f58","MirrorCode 把长程编程拖进可测量区间：Opus 4.7 重写 6 万行 Pkl，AI 编码能力一年翻倍","mirrorcode-long-horizon-opus-pkl-56pct","2026-06-28T02:03:00+00:00",{"id":65,"title":66,"news_slug":67,"published_at":68},"30c32de0-5d4e-4c7b-b0d7-35b27f776e4f","DeepSWE 接管 Coding Agent 评测：SWE-Bench Pro 32% 误判如何被基准审计撕开","deepswe-datacurve-coding-agent-32pct-misjudge","2026-06-16T22:30:00+00:00",{"id":70,"title":71,"news_slug":72,"published_at":73},"5c23c8b7-693f-415d-a255-beea9b465f67","2026年LLM评估风向变了：MMLU不再是主角，SWE-Bench登基","2026-llm-benchmark-swe-bench-king-mmlu","2026-05-30T19:03:00+00:00",{"id":75,"title":76,"news_slug":77,"published_at":78},"b95c3074-f3f1-4473-8df5-0f625c332a8d","AgentEscapeBench：美团+复旦推出工具推理评测新基准，揭示大模型Agent深层依赖短板","agentescapebench-meituan-fudan-dag-270-tasks","2026-05-12T07:01:00+00:00",{"id":80,"title":81,"news_slug":82,"published_at":83},"aca19c8b-6ba8-49d8-b761-fb812fb18a77","Kimi K2.6 在AI编程挑战赛中夺冠：开源模型展现长时推理优势","kimi-k2-6-programming-challenge-22-pt-beats-gpt-5-5","2026-05-04T01:01:00+00:00"]