推荐理由：多数团队会堆叠路由器、微调与多模型，OpenAI 的数据智能体却只用单模型加 13 个工具跑通 90,000 张表，形成可对比的工程取舍。
Agent 智能体
全部主题让模型自主规划、调用工具、完成多步任务的技术方向——从 Claude Code、Manus 到各家 Agent 框架与评测基准的全部动态。
当前仅显示精选新闻最新精选
第 641–660 条 · 共 820 条@alexxubyte@alexxubyte精选AI 评分6666 @swyx@swyx精选AI 评分7474 
引用OpenAI (@OpenAI)@OpenAIWe’re making Codex more useful for your work by expanding plugins beyond individual tools. These plugins turn Codex into a specialist for a specific role with a single install, no coding required. Codex can access 62 popular apps and 110 skills for work across sales, data analytics, creative production, product design, and public equity investing. openai.com/index/codex-for-e…
推荐理由：Codex 借插件把多款工具打包成角色化安装包，读者可据此了解其能力从编码向销售、数据分析等场景延伸。
@berryxia@berryxia精选AI 评分6565 
引用Google DeepMind (@GoogleDeepMind)@GoogleDeepMindWe believe AI can be a dedicated research partner to help discover the next breakthrough. Enter Co-Scientist: our latest Gemini-based multi-agent system that can generate, debate and evolve novel hypotheses for complex scientific problems 🧵 Video
推荐理由：原文说明了 Co-Scientist 的多 Agent 假设生成与验证流程，并指出该能力已向个人研究者开放。
@eliebakouch@eliebakouch精选AI 评分7373 
引用Mustafa Suleyman (@mustafasuleyman)@mustafasuleymanSuper excited to announce seven new world-class MAI models today. They represent what we consider a new era in AI designed to keep you in control and on the frontier. First is our text foundation model, MAI-Thinking-1, exceptionally strong on reasoning and SWE tasks. - It’s a 35B active parameter MoE with a 256K context window. Independent human raters on Surge prefer it for overall quality in blind side-by-sides versus Sonnet 4.6, and it’s achieved 97% on AIME 2025, the key measure of its general-purpose reasoning abilities. - It's at 53% on SWE Bench Pro, placing it right alongside Opus 4.6 on one of the toughest coding benchmarks. - And since we co-designed our models with our own silicon, MAI-Thinking-1 is optimized on our MAIA 200 chip. Benchmarking head-to-head against the GB200, we see 30% better performance per dollar as well as a 1.4x performance-per-watt gain when running our MAI models on the MAIA 200 end-to-end. Next is MAI-Image-2.5 and its Flash variant. Two super strong models now at #2 on the leaderboards, surpassing the score of Nano Banana 2 on image editing. Last for now is MAI-Code-1-Flash, our new inference efficient coding model, especially tuned for VS Code and GitHub Copilot CLI. - Code-1-Flash achieves 51% on SWE Bench Pro, despite having just 5B parameters, putting it closer to Haiku in size but cheaper in cost. All of this is the foundation for Microsoft Frontier Tuning. It lets you customize our models to create custom, company-specific agents that only you control. You can make our model, your model. Your data. Your agents. Your moat. Early adopters are already seeing a difference. When we tuned our models for McKinsey’s tasks, MAI delivered the highest win rate, outperforming GPT-5.5 on quality, while being 10x lower on cost. Also really excited to be collaborating with the amazing team at Mayo Clinic to jointly train a new frontier AI model for healthcare. Our announcements today mark another milestone on the road to humanist superintelligence. You can learn more and about our other new models in our latest blog: microsoft.ai/news/building-a…
推荐理由：作者逐点点评微软 MAI 技术报告，读者可了解其无合成数据与蒸馏的训练取舍及 scaling ladder 的公开细节。
Claude Blog精选AI 评分6161 Anthropic 如何用 Claude 实现自助式数据分析
Anthropic 在官方博客中介绍，其内部 95% 的业务分析查询已交由 Claude 自动完成，整体准确率约 95%。文章给出智能体分析栈的四层结构，即数据基础、真相来源、技能和验证，并归纳出概念与实体歧义、数据陈旧、检索失败三类主要错误来源。Anthropic 称未加技能时 Claude 回答分析问题的准确率不足 21%，加上技能后稳定超过 95%，部分领域可达约 99%。
推荐理由：原文详解 Anthropic 把业务分析交给 Claude 的方法，归纳三类失败模式与分层治理思路，可迁移到其他数据团队。
Claude Blog精选AI 评分7979 Anthropic 分享 Claude Code 内部使用 skills 的经验
Anthropic 的 Claude Code 团队分享了内部数百个 skills 的使用经验，并把它们归纳为九大类别，包括库与 API 参考、产品验证、数据获取分析、业务自动化、代码脚手架、代码质量审查、CI/CD 部署、运行手册和基础设施运维。
推荐理由：Anthropic 把内部数百个 skills 归纳为九类并给出编写要点，可供团队搭建自己的 skills 库参考。
Claude Blog精选AI 评分6363 Anthropic Claude Code 团队如何重构 AI 原生工程组织的流程
Anthropic Claude Code 团队负责人复盘了智能体编码成为默认工作方式后对工程流程的重写：规划从六个月路线图改成 just-in-time，遇到问题先问 Claude 并追问能否自动化，代码评审由 Claude 处理风格、缺陷和测试，人只在法律风险、信任边界与安全代码、产品判断等需要领域专长处介入。
推荐理由：Claude Code 团队把规划、上下文获取、评审与分工按智能体编码重写，并给出可对照的前后变化与跟踪指标。
@OpenAI@openai精选AI 评分7676 推荐理由：原文给出可一次安装的角色化插件、接入应用与技能数量，读者可据此判断 Codex 能覆盖哪些岗位工作。
@berryxia@berryxia精选AI 评分7070 引用Vaibhav (VB) Srivastav (@reach_vb)@reach_vbWe just released the Codex Python SDK 🔥 You can now embed Codex directly into your Python apps and workflows! > Start threads > Run turns > Stream progress > Resume sessions > Pass images > Control sandbox access All whilst reusing your existing Codex auth. pip install openai-codex Go build with it!!
推荐理由：SDK 把 Codex 从浏览器里的 IDE 变成可嵌进脚本的可编程基础设施，并复用现有认证。
@vista8@vista8精选AI 评分7070 引用Vaibhav (VB) Srivastav (@reach_vb)@reach_vbWe just released the Codex Python SDK 🔥 You can now embed Codex directly into your Python apps and workflows! > Start threads > Run turns > Stream progress > Resume sessions > Pass images > Control sandbox access All whilst reusing your existing Codex auth. pip install openai-codex Go build with it!!
推荐理由：原文给出 Codex Python SDK 的安装命令与复用登录态方式，读者可据此把它嵌入 Python 应用和工作流。
Microsoft AI News精选AI 评分6060 Microsoft AI 阐述医疗领域布局：与 Mayo Clinic 合作开发医疗前沿模型并开放 Copilot Health 预览
Microsoft AI 发文阐述其在医疗领域的产品与研究方向，称旗下消费产品每天处理数千万条健康问题，对 50 万条 Copilot 对话的分析显示用户依赖 AI 做症状解读、检查结果解释和就医导航。
推荐理由：微软官方阐述其在医疗领域的价值落地，读者可以据此了解 MAI DxO、Mayo Clinic 合作与 Copilot Health 的进展全貌。
Hugging Face Blog精选AI 评分7272 H 公司发布 Holo3.1 计算机使用智能体模型，推出本地量化权重
H 公司发布 Holo3.1 计算机使用智能体模型家族，提供 0.8B、4B、9B 和 35B-A3B 四种尺寸，并首次推出 FP8、Q4 GGUF 和 NVFP4 量化权重。
推荐理由：Holo3.1 在移动端与跨框架适配上补齐短板，量化权重把计算机使用智能体带到本地设备。
@MiniMax_AI@minimax_ai精选AI 评分7373 
引用MiniMax (official) (@MiniMax_AI)@MiniMax_AIIntroducing MiniMax M3: The First Open-Weights Model to Combine Three Frontier Capabilities - Coding & Agentic Frontier: 59.0% SWE-Bench Pro, 66.0% Terminal Bench 2.1, 34.8% SWE-fficiency, 28.8% KernelBench Hard, 74.2% MCP Atlas - MiniMax Sparse Attention scales context to 1M - Natively Multimodal from Step Zero API: platform.minimax.io Token Plan: platform.minimax.io/subscrib… 🚀New! MiniMax Code: code.minimax.io Weights & Tech Report in ~10 Days
推荐理由：官方给出 M3 在编码与智能体基准上的具体分数和 1M 上下文，读者可据此对照同类开源模型的能力边界。
@MiniMax_AI@minimax_ai精选AI 评分6868 
引用MiniMax (official) (@MiniMax_AI)@MiniMax_AIIntroducing MiniMax M3: The First Open-Weights Model to Combine Three Frontier Capabilities - Coding & Agentic Frontier: 59.0% SWE-Bench Pro, 66.0% Terminal Bench 2.1, 34.8% SWE-fficiency, 28.8% KernelBench Hard, 74.2% MCP Atlas - MiniMax Sparse Attention scales context to 1M - Natively Multimodal from Step Zero API: platform.minimax.io Token Plan: platform.minimax.io/subscrib… 🚀New! MiniMax Code: code.minimax.io Weights & Tech Report in ~10 Days
推荐理由：原文列出 M3 的编码与 Agent 基准成绩及 1M 上下文能力，可对照开源模型的当前水位。
Claude Blog精选AI 评分7676 Claude Code 发布动态工作流，可为每个任务生成专属 harness
Anthropic 在 Claude Code 中发布动态工作流，Claude 可按当前任务即时编写专属 harness，并与 Claude Opus 4.8 配合使用。工作流通过 agent() 生成子智能体，用 parallel() 和 pipeline() 组合编排，可保存为可复用工作流或通过 skill 分享。文中提到动态工作流消耗更多 token，更适合复杂高价值任务。
推荐理由：文章给出动态工作流的六种编排模式和适用边界，读者可据此判断何时该用多智能体并行而非单上下文。
@alibaba_cloud@alibaba_cloud精选AI 评分7474 
推荐理由：官方列出多模态智能体的能力清单与 API 入口，读者可据此判断它在自身工作流中的可用性。
@kimmonismus@kimmonismus精选AI 评分6565 引用Qwen (@Alibaba_Qwen)@Alibaba_Qwen👏👏 Introducing Qwen3.7-Plus — a multimodal agent model that unifies vision and language into one versatile agent foundation. ✅ Multimodal interactive hybrid agent: unified GUI & CLI operation across visual and text tasks ✅ Versatile coding agent & productivity assistant with full-modality input ✅ Visual Agent: perception, reasoning, grounding, and search-augmented QA ✅ Cross-harness generalization across diverse agent frameworks One model. Sees, thinks, codes, acts.🙌🙌 Now available via API on Alibaba Cloud Model Studio. Try it — let us know what you build.😎 🔗🔗⬇️⬇️ Blog：qwen.ai/blog?id=qwen3.7-plus Qwen Studio：chat.qwen.ai/?models=qwen3.7… API：modelstudio.console.alibabac…
推荐理由：官方把 Qwen3.7-Plus 与 GPT-5.4、Opus 4.6 并列对比，读者可据此观察多模态智能体模型的竞争位置。
@kimmonismus@kimmonismus精选AI 评分6767 
引用NVIDIA Newsroom (@nvidianewsroom)@nvidianewsroomIntroducing NVIDIA DGX Station for Windows, the world's most powerful deskside AI supercomputer with Windows powered by NVIDIA GB300. ✅ Run frontier AI models with up to 1 trillion parameters locally ✅ Build and run secure AI agents on Windows with NVIDIA OpenShell ✅ Built by @ASUS, @Dell, @GIGABYTE, @HP, @msigaming, and @Supermicro #NVIDIAGTC nvda.ws/3RIkjpc
推荐理由：原文列出 GB300 桌面超级芯片的完整规格与出货时间，可据此判断本地运行前沿模型与智能体的硬件门槛。
@kilocode@kilocode精选AI 评分6767 引用MiniMax (official) (@MiniMax_AI)@MiniMax_AIIntroducing MiniMax M3: The First Open-Weights Model to Combine Three Frontier Capabilities - Coding & Agentic Frontier: 59.0% SWE-Bench Pro, 66.0% Terminal Bench 2.1, 34.8% SWE-fficiency, 28.8% KernelBench Hard, 74.2% MCP Atlas - MiniMax Sparse Attention scales context to 1M - Natively Multimodal from Step Zero API: platform.minimax.io Token Plan: platform.minimax.io/subscrib… 🚀New! MiniMax Code: code.minimax.io Weights & Tech Report in ~10 Days
推荐理由：官方给出 M3 的编程与 Agent 基准数据及 1M 上下文能力，可据此观察开源权重模型的前沿能力边界。
@kimmonismus@kimmonismus精选AI 评分7474 
引用MiniMax (official) (@MiniMax_AI)@MiniMax_AIIntroducing MiniMax M3: The First Open-Weights Model to Combine Three Frontier Capabilities - Coding & Agentic Frontier: 59.0% SWE-Bench Pro, 66.0% Terminal Bench 2.1, 34.8% SWE-fficiency, 28.8% KernelBench Hard, 74.2% MCP Atlas - MiniMax Sparse Attention scales context to 1M - Natively Multimodal from Step Zero API: platform.minimax.io Token Plan: platform.minimax.io/subscrib… 🚀New! MiniMax Code: code.minimax.io Weights & Tech Report in ~10 Days
推荐理由：原文给出了 M3 与 GPT-5.5、Opus 4.7 的基准对比和成本差距，读者可据此判断开放权重模型的能力位次。