AI 资讯 · 2026-09-22

9月22日，在2026年云栖大会上，千问宣布将加速打造 Personal Agent。千问产品负责人郑嗣寿表示，自去年上线以来，让每个人拥有自己的 AI 助理，始终是千问不变的目标。依托Qwen 3.8系列模型强大的Agentic能力，千问正在构建全新的Personal Agent形态，为 3 亿用户提供持续、个性化的智能服务。当天，千问公布了多项围绕 Personal Agent 的阶段性进展：在用户授权下接入健康、运动等个人数据，为健康咨询和运动规划提供更具针对性的分析；接入海量专业理财数据及机构智能体，并支持部分用户连接个人持仓；同时，千问开放平台覆盖领域已扩大至20多个，申请入驻伙伴超过1000家。目前，千问正将Agent能力逐步落地到日常生活、运动健康、工作学习等更多场景中。在理财场景中，用户可在对话中告诉千问自己的理财目标，千问将结合实时行情、专业机构信息及用户授权连接的持仓数据，提供更贴近个人情况的分析与建议。千问还将金融机构的专业研究能力封装为Skill，让普通用户也能获得专业的决策支持。针对理财场景的复杂任务，千问进行了专项 Harness 优化，回复效果提升20%。

Rabbit, the company behind the underwhelming R1 device, is rolling out a standalone AI agent that you don't need its hardware to use, as reported earlier by Wired. The startup says its new OS3 "agentic operating system" runs in the cloud but operates locally across Windows, Mac, and Linux devices. According to Rabbit, you can […]

一场两小时的客户拜访，双方说的话加起来上万字。但回到工位，销售能凭记忆敲进CRM的，通常不到三百字。剩下的去哪了？跟着出租车一起走了。这样的对话在企业里每天都在发生。会议室里的争论、客户办公室里随口的一句提醒、老师傅肚里的经验，信息密度最高的这些话，几乎都发生在电脑屏幕之外，从来没有进入过任何一套企业系统。另一边，已经沉淀在文档、群聊和系统里的数据，也常常因为没做结构化，或者体量太大，塞不进上下文窗口，Agent根本调用不了。9月22日云栖大会，千问办公把两件事放到了一起解决：发布了面向企业级Agent的基础设施“企业上下文”（Enterprise Context），以及Agent硬件QwenNote A2。简单说，Enterprise Context把散落在群聊、文档、知识库和业务系统里的企业信息压缩、结构化，随业务变化持续更新，变成Agent能随时调用的知识。难的地方在“给多少”。一家公司积累多年的资料动辄几十万条，但员工开始一项具体任务时，需要的从来不是全部，而是跟眼下这件事最相关的那一小块。企业上下文要做的就是这件事，每次只把最相关的那一小块递过去，再用内置的专属模型把Toke

当经典算法遇到 LLM，用概率预判代替盲目试错，或许才是架构演进的真正解法。    作者丨张   璐    编辑丨马晓宁                                                                                                       过去，智能体的技能（Skill）大多依赖人类专家手动撰写，不仅耗时费力，更难以覆盖海量复杂的真实场景。为此，借助大模型“自动化生成与优化技能”，成了让 Agent 迈向自主进化的新范式。然而，摆脱了人工编写的束缚，却撞上了高昂的 Token 账单。要验证一个自动生成的技能到底管不管用，必须让智能体带入真实任务中落地执行。每一次试错，都是实打实的 Token 消耗与多轮交互的时间成本。更棘手的是，许多候选技能在刚生成时就自带缺陷。但在现有框架下，我们缺乏“事前筛选”的能力，只能任由大量预算被这些低质候选无意义地吃掉。同时，后续的精炼流程也在频繁调用高阶 LLM，哪怕缺乏有效的新证据，这种“反思—重写”的流水线依然在机械重复。评估贵，精炼也贵，两笔账算下来，技能优化的效

OpenAI is launching two new models, which it says are cut from the same cloth as Astra.

At TechCrunch Disrupt 2026, five sessions across the AI Stage and Real World AI Stage cover AI safety, featuring leaders from Anthropic, Nvidia, AWS, Waabi, and more. Register before September 25 to save up to $200.

British Columbia sues OpenAI, demands Tumbler Ridge shooter’s ChatGPT logs.

Learn how GPT-6 improves prompt caching with higher cache hit rates, new diagnostics, explicit breakpoints, and controls that reduce latency and costs.

The frontier AI model race has entered its comparison shopping phase.

It’s been a busy few months for AI hype. At the end of April, Anthropic claimed that its model Claude Mythos is better at finding software vulnerabilities than most security experts. Then we had the OpenAI–Hugging Face hacking incident, after which Anthropic (proudly) and Meta (reluctantly) disclosed similar incidents involving their models. This was followed…

OpenAI outlines priorities and principles for rigorous, secure, and independent third-party AI safety assessments of frontier models and safeguards.

Have you ever read a paper in Science or Nature and thought, “Man, that research was so cool. I wish I could try that method on my own data,” only to spend a week wrestling with someone else’s undocumented repo, broken dependencies, and half-finished readme.txt?Well, now you can, more or less.Say hello to Paper2Agent, a new open-source framework that transforms academic reports into interactive AI agents you can talk to. Give it a paper, along with the accompanying codebase, data, or other suppl

Two years after trying to sidestep mobile apps with dedicated AI hardware, Rabbit is launching OS3, a cross-platform agent that lives on the screens you already use.

Meta has issued a patch for its Muse macOS app following the discovery of a zero-day vulnerability that could allow someone to take control of the AI agent. The bug found by security researcher Patrick Wardle utilized an undocumented Muse setting that enabled potential attackers running local code to redirect transcription processing from Meta's servers […]

In the hope of uncovering new details about ancient life, researchers have developed a large language model that fills in the gaps in papyrus fragments.

GPT‑6 Astra allowed Parallel’s agents to research and synthesize labor-market data in half the time and at half the cost vs. prior models.

EvilTokens provided an end-to-end platform that makes mass compromises faster and easier.