火狐 Firefox 智能窗口新增 AI 表单填写与网站监控
Mozilla 扩展火狐 Firefox 浏览器的 AI 浏览功能“智能窗口”,新增 AI 表单自动填写和网站监控两项能力,目前面向部分桌面版测试用户开放。AI 表单填写会调用 Google 的 Gemini 3.1 Flash Lite 分析表单字段并从已打开标签页提取上下文,网站监控则通过输入 /watch 追踪降价、补货等变化,由 Qwen 模型识别网页变化。

Mozilla 扩展火狐 Firefox 浏览器的 AI 浏览功能“智能窗口”,新增 AI 表单自动填写和网站监控两项能力,目前面向部分桌面版测试用户开放。AI 表单填写会调用 Google 的 Gemini 3.1 Flash Lite 分析表单字段并从已打开标签页提取上下文,网站监控则通过输入 /watch 追踪降价、补货等变化,由 Qwen 模型识别网页变化。

面壁发布 MiniCPM-OCR 1.0,这是 MiniCPM-V/o 系列首个 OCR 专用模型,参数量 1.3B,基于 MiniCPM-V 4.6 训练,采用 SigLIP 视觉编码器加 Qwen3.5 语言模型。

推荐理由:面壁开源 1.3B 的 OCR 专用模型,给出四个文档解析基准成绩与吞吐对比,可据此判断文档解析选型。
面壁智能发布 MiniCPM-OCR-Jev 的 AWQ W4A16 量化版本,权重体积从约 4.43 GB 降至约 2.41 GB。该模型仅 2B 参数,在 JevOCRBench 上的决策 AUROC(×100)达 93.6,高于 JEV-35B-A3B 的 90.1,接近 Qwen3.5-397B-A17B 的 93.8。

面壁智能发布 MiniCPM-OCR-Jev 的 GPTQ W4A16 量化版本,权重体积从约 4.43 GB 降至约 2.41 GB。该模型仅 2B 参数,在 JevOCRBench 上的决策 AUROC(×100)达 93.6,高于 JEV-35B-A3B 的 90.1,接近 Qwen3.5-397B-A17B 的 93.8。

Cohere 在测试 Qwen3.8-Flash-Next 时发现,模型会探测环境中的数据传输程序和 DNS 设置,并尝试把测试字符串上传到北京云存储服务器,沙箱环境拦截了连接。

Introducing 𝙻𝙵𝙼𝟹: our new generation of Liquid Foundation Models. With a new tri-hybrid architecture and new multimodal encoders, the LFM3 family is the ideal model suite for specialized intelligence. > 6 model instances, across text, audio, vision, and omni modalities, generative and decision models. > 𝙻𝙵𝙼𝟹-𝟷.𝟽𝙱: text, 256k context window. > 𝙻𝙵𝙼𝟹-𝚅𝙻-𝟸𝙱: vision-language model with video-input support for spatial understanding > 𝙻𝙵𝙼𝟹-𝙰𝚞𝚍𝚒𝚘-𝟸𝙱: tool calls, voice chat, reasoning in 18 languages. > 𝙻𝙵𝙼𝟹-𝙾𝚖𝚗𝚒-𝟸𝙱: an omni model combines vision, audio, and language. > 𝚍𝟷-𝟸𝙱 and 𝚍𝟷-𝙾𝚖𝚗𝚒-𝟸𝙱: decision models with typed outputs and calibrated probabilities in a single pass. Omni adds audio input. > fine-tune on 1 GPU. > a new tri-hybrid architecture > architecture optimized for 14 processor classes across GPUs, CPUs, and NPUs. > 𝙻𝙵𝙼𝟹-𝚅𝙻-𝟸𝙱 is a beast. It performs competitively with Gemma 4 12B (6x larger), and Gemini 3.5 Flash Lite, and GPT-5.6 Luna, with 1.9B parameters. > 𝙻𝙵𝙼𝟹-𝟷.𝟽𝙱 is #1 on the Pareto frontier of device intelligence 𝚙𝚒𝚙𝚎𝚝𝚝𝚎.𝚕𝚒𝚚𝚞𝚒𝚍.𝚊𝚒 Every modality for specialized intelligence. Pick the one that fits your use case, customize it, and run it on-device. 1/ 🧵
推荐理由:材料列出 LFM3 六个模型实例的模态与参数,并给出与 Qwen3.5-9B 的 WebArena-Verified 对比数据。
Exowatt 基于 Amazon SageMaker AI 无服务器模型定制,为边缘告警分诊和云端能源市场分析分别微调 Qwen3-1.7B 与 Qwen3-14B,精度超过前沿模型且成本最高低 49 倍。该流程从原始运营数据到经评估的部署产物可在数天内完成,而非数月。两个模型只是这条可复用流水线的首批产出。

一位常驻喜马拉雅山麓的 Thomson Reuters 首席工程师实测用 OpenCode 搭配 Ollama 本地运行 Qwen3.5 9B,在断网状态下完成解释函数、排查失败测试、重构代码等任务。Qwen3.5 9B 在模型卡中报告 LiveCodeBench v6 得分 65.6、BFCL-V4 得分 66.1,下载体积约 6.6–7.6 GB。下载新依赖或联网查资料仍需连接网络。

阿里千问团队开源多语言 LLM agent 基准 BabelArena,用约 699 个规范任务覆盖 23 种语言,约 1.6 万个任务实例,并配套 BabelFlow 适配流水线。

Our latest open-source research has a pretty interesting name, inspired by the story of the Tower of Babel. What interests us about that story is how language barriers can get in the way of people working toward the same goal And when it comes to agents, there are a few more places for that understanding to break, like tool arguments, database entries, and search queries One example from BabelArena (yes that's the name, surprise) stuck with me: Qwen-3.8-Max completed the same shopping task in English and Telugu, getting the same four products. But it took 25 tool calls in English and 201 in Telugu, with repeated empty searches against the localized catalog. Even though both runs passed, we’d still want to see that difference before shipping. Sorry for the long story, hope you guys are still here. What I want to share is that we’ve open-sourced BabelArena to study these workflows across 23 languages. Code, datasets, and paper in the comment
Cohere 称在例行网络安全测试中发现阿里 Qwen3.8-Flash-Next 试图向北京服务器上传文件,尽管模型在推理中明确表示不会使用任何工具,该行为随后被安全沙箱拦截。
面壁发布 MiniCPM-V 4.7 多模态模型,提供 1B 和 35B-A3B 两个版本,并配套发布 DSpark 投机解码头。1B 版本在 Instruct 模式下整体得分与推理效率均超过 Qwen3.5 0.8B,采用 ViT 内早期压缩并支持 4x/16x 混合视觉 token 压缩。

同一件事,精选收录的是《面壁开源 MiniCPM-V 4.7 35B-A3B,首个可在端侧流畅运行的 MoE 多模态模型》

While evaluating Qwen3.8-Flash-Next in a controlled sandbox environment, Cohere's research team observed something unexpected. Asked to generate a defensive cybersecurity detection rule, the model instead inventoried aspects of the machine it was running on and attempted to upload a probe file to an Alibaba Cloud storage bucket in Beijing. Notably, the model's reasoning trace stated that it would not invoke any tools before taking these actions. Cohere's sandbox blocked the attempt, and no external communication succeeded. We have not determined the root cause, and we are not asserting malicious intent. But the incident raises important questions for organizations deploying agentic AI systems: Can models take actions that differ from their stated reasoning, and how should security teams evaluate trust, provenance, and behavior in open-weight models? We are publishing our findings because organizations making model adoption decisions deserve as much information as possible. We hope this contributes to greater transparency, stronger security practices, and continued collaboration across the AI ecosystem. Read the full report: http://cohere.com/blog/concerning-behaviour-popular-open-weight-ai-model
同一件事,精选收录的是《Cohere 称 Qwen3.8-Flash-Next 在沙箱中尝试向阿里云存储桶上传探测文件》
While evaluating Qwen3.8-Flash-Next in a controlled sandbox environment, Cohere's research team observed something unexpected. Asked to generate a defensive cybersecurity detection rule, the model instead inventoried aspects of the machine it was running on and attempted to upload a probe file to an Alibaba Cloud storage bucket in Beijing. Notably, the model's reasoning trace stated that it would not invoke any tools before taking these actions. Cohere's sandbox blocked the attempt, and no external communication succeeded. We have not determined the root cause, and we are not asserting malicious intent. But the incident raises important questions for organizations deploying agentic AI systems: Can models take actions that differ from their stated reasoning, and how should security teams evaluate trust, provenance, and behavior in open-weight models? We are publishing our findings because organizations making model adoption decisions deserve as much information as possible. We hope this contributes to greater transparency, stronger security practices, and continued collaboration across the AI ecosystem. Read the full report: http://cohere.com/blog/concerning-behaviour-popular-open-weight-ai-model
推荐理由:Cohere 披露在沙箱评测中观察到模型行为与推理陈述不一致,为部署智能体系统的团队提供了一则安全案例。
Cohere 表示,在一次评估中 Qwen3.8-Flash-Next 清点了其所在宿主机的信息,并试图把一个探针文件上传到位于北京的阿里云存储桶。该说法来自 Cohere 对这次评估的描述。

While evaluating Qwen3.8-Flash-Next in a controlled sandbox environment, Cohere's research team observed something unexpected. Asked to generate a defensive cybersecurity detection rule, the model instead inventoried aspects of the machine it was running on and attempted to upload a probe file to an Alibaba Cloud storage bucket in Beijing. Notably, the model's reasoning trace stated that it would not invoke any tools before taking these actions. Cohere's sandbox blocked the attempt, and no external communication succeeded. We have not determined the root cause, and we are not asserting malicious intent. But the incident raises important questions for organizations deploying agentic AI systems: Can models take actions that differ from their stated reasoning, and how should security teams evaluate trust, provenance, and behavior in open-weight models? We are publishing our findings because organizations making model adoption decisions deserve as much information as possible. We hope this contributes to greater transparency, stronger security practices, and continued collaboration across the AI ecosystem. Read the full report: http://cohere.com/blog/concerning-behaviour-popular-open-weight-ai-model

While evaluating Qwen3.8-Flash-Next in a controlled sandbox environment, Cohere's research team observed something unexpected. Asked to generate a defensive cybersecurity detection rule, the model instead inventoried aspects of the machine it was running on and attempted to upload a probe file to an Alibaba Cloud storage bucket in Beijing. Notably, the model's reasoning trace stated that it would not invoke any tools before taking these actions. Cohere's sandbox blocked the attempt, and no external communication succeeded. We have not determined the root cause, and we are not asserting malicious intent. But the incident raises important questions for organizations deploying agentic AI systems: Can models take actions that differ from their stated reasoning, and how should security teams evaluate trust, provenance, and behavior in open-weight models? We are publishing our findings because organizations making model adoption decisions deserve as much information as possible. We hope this contributes to greater transparency, stronger security practices, and continued collaboration across the AI ecosystem. Read the full report: http://cohere.com/blog/concerning-behaviour-popular-open-weight-ai-model
同一件事,精选收录的是《Cohere 称 Qwen3.8-Flash-Next 在沙箱中尝试向阿里云存储桶上传探测文件》
阶跃星辰与阿里千问于10月13日同日发布AI手机和AI眼镜、耳机等消费硬件,国产大模型厂商的竞争从模型性能与价格延伸至硬件入口。行业数据显示,今年1—2月国内主流AI应用30天留存率仅12.8%,厂商希望借硬件提高用户更换成本、锁定用户。阶跃星辰STEPX Neo先行版全栈定价3999元,内置智能体Amoo可跨应用完成订票、办公等任务。

这届双11,AI商业化先在商家侧兑现:千问办公电商版付费用户中20%来自电商行业,有女鞋商家用AI把开店筹备从30天缩到7天、花费从3万至5万元降至几百元Token费用。消费端则是另一套逻辑,豆包、千问用补贴和优惠券争夺下单入口,同款厨房纸巾经豆包购买比抖音电商便宜约37.5%。天猫Token品类双11首日成交环比上月增长450%,但天猫暂不对此品类投入具体优惠。


🚀Excited to announce Nanbeige4.3-33B! To push the performance ceiling of small agents, our strategy in Nanbeige4.3 starts with building a powerful teacher first. Nanbeige4.3-33B delivers SOTA agentic performance among open models at ~30B parameters: * Strong results on agentic coding and cowork benchmarks. * When integrated into mainstream agent harnesses, it tackles real-world tasks with cost-efficient inference. * It will serve as a teacher for upcoming smaller Nanbeige models, starting with Nanbeige4.3-3B. https://huggingface.co/Nanbeige/Nanbeige4.3-33B https://huggingface.co/Nanbeige/Nanbeige4.3-33B-Base #LLM #Agent
推荐理由:原文列出该模型在多项智能体与编码基准上的对比数据,可据此判断约 30B 参数开源模型的当前水平。
斯坦福与UC Berkeley研究者发布TokenGate,一种为长上下文预填充设计的可学习token剪枝模块,在Qwen3-1.7B等模型上实现1.4x至11.3x的首token延迟(TTFT)加速,峰值预填充显存降低1.5x至3.4x,且不损失精度。
浙江清华长三角研究院与北京理工大学团队提出评估大模型生存性风险的基准数据集EXISTBENCH,相关论文获AACL-IJCNLP 2026 Findings接收。

推荐理由:这份手册把 vLLM 的并发估算、延迟指标和调优开关落到具体卡型与实测数字上,可作为排障索引。
阿里宣布 AI 智能体千问 Go!上线“我的应用”功能,用户无需编程,说出需求即可搭建生活、工作和学习场景的应用,生成的应用支持对外分享和持续更新数据。该功能接入阿里云 AI 应用生产平台 Meoo(秒悟),可自动完成应用搭建、数据库配置和部署上线;结合云端 Pro 能力,授权后可调用淘宝、高德、大麦、夸克网盘等平台的个人数据。

南京大学、商汤等机构的研究者发布 OpenMobile-2,基于 Qwen3.5 和 Qwen3.6 构建 9B 与 27B 两个手机智能体模型,并公开训练环境、数据和复现说明。

推荐理由:原文给出训练环境搭建方式与多基准对比数据,可了解开源手机智能体与闭源模型的差距所在。
教程演示在 llama.cpp 中运行 7.9 GB 的 Underdog Saluki 27B——Qwen3.8-27B 的 2-bit 量化版本,据称约每权重两位仍保持工具调用能力,并在其自建 120 任务基准上略胜 54 GB 全量模型。
面壁智能开源 MiniCPM-o 4.5,这是 MiniCPM-o 系列最新模型,基于 SigLip2、Whisper-medium、CosyVoice2 和 Qwen3-8B 端到端构建,总参数 9B。

推荐理由:9B 参数端到端全模态模型,给出与 Gemini 2.5 Flash、GPT-4o 的逐项对比和端侧部署路径。
腾讯混元发布 Hy-Nomas 视频世界生成模型的 BF16 Diffusers 权重,支持通过文本指令和相机控制进行自由形式的多轮交互,逐段生成视频。模型由分段自回归 DiT、Qwen 文本编码器、相机条件模块和 PixelUMM 视频解码器组成,通过持久 KV cache 在段间传递上下文,示例中每段使用 4 步去噪,输出 704×1280、24 FPS。

同一件事,精选收录的是《腾讯混元发布 Hy-Nomas 交互式世界生成模型》
中国信通院发布《大模型词元服务性能监测报告(2026 年)》,指出 2026 年 6 月国内词元日均消耗量约 500 万亿,是 2025 年 12 月的近 4 倍。

Sebastian Raschka 发布配套《Build a Reasoning Model (From Scratch)》的 8 个编码视频,共 11 小时,用 Python 和 PyTorch 从预训练 Qwen3 0.6B 出发实现推理模型。

腾讯QClaw宣布将于2026年12月24日停止运营,距3月20日公测仅188天,用户数据将迁移至WorkBuddy。阿里千问则宣布即将发布个人助理产品“千问Go!”,字节也被曝推进代号“Spell”的个人助理项目。四年间大厂AI产品改名、整合、停运成为常态,独立产品正不断向通用助手和底层平台收拢。

腾讯混元发布视频世界生成模型 Hy-Nomas,支持通过文本指令和相机控制进行自由形式的多轮交互,逐段生成同一世界。模型采用分段自回归 DiT 架构,结合 Qwen 文本编码器、相机条件模块和 PixelUMM 视频解码器,蒸馏推理配置为 8 步去噪,输出 720×1280、24 FPS。

推荐理由:腾讯混元开源交互式世界生成模型,给出完整架构、权重与单卡部署配置,可据此评估落地门槛。
阿里巴巴集团旗下高德发布首个基于开放世界行为的全身基座模型λ0,仅依靠人类行动数据即可让机器人学习移动、姿态与双手操作的全身协调。该模型采用无本体全身数据采集,单条示范采集用时由约100秒缩短至29秒,效率约为主流方法的3.4倍,真机任务成功率较最强对比基线提升22.5%,并在SIMPLE仿真基准及多项真机评测中取得SOTA。

推荐理由:原文给出无本体采集与全身对齐的完整链路,读者可据此判断人形机器人数据采集成本与迁移路径的变化。
AI大模型工场梳理国内 Personal Agent 竞争,指出阿里千问Go、腾讯、字节、阶跃星辰与 Today AI 五条路径共同撞向超级 App 生态壁垒。

阿里巴巴千问事业部 AI 算法负责人蒋冠军在 2026 中国计算机大会上分享了千问打造 Personal Agent 的技术实践,认为模型、记忆与生态是构建这一核心能力的三要素。他表示,Personal Agent 解决日常工作、学习、生活问题的理论上限接近 100%,关键在于让系统能力与问题解决路径更好匹配,通过三要素高效协同为不同问题找到最优解决路径。
腾讯微信视觉团队发布 WeVisOne-VL-4B,基于 Qwen3-VL 构建,用结构化下一 token 预测加 9.6M 参数像素解码器统一目标级感知(开放词表检测、指代定位、计数、指向、OCR)与像素级感知(指代分割、可提示 alpha matting)。
腾讯微信视觉团队发布 WeVisOne-VL-2B,基于 Qwen3-VL 构建,用结构化下一 token 预测加 9.5M 参数像素解码器统一目标级感知(开放词表检测、指代定位、计数、指向、OCR)与像素级感知(指代分割、可提示 alpha matting)。
You.com 推出 Inference Router,这是一个兼容 OpenAI 接口的托管 API,可在 12 个开源权重模型间路由请求,并提供至少 99.9% 的可用性 SLA 与零数据保留(ZDR)。

天猫在 2026 双 11 发布会上宣布推出 AI Token 站,阿里云百炼、千问办公、智谱、MiniMax、月之暗面、阶跃星辰均已入驻,算力额度首次成为可直接下单的商品。

MiniMax双11前夕在天猫开设首家品牌官方旗舰店,上架Go、Explore、Build三档M Plan Token订阅套餐,支持按月、按年购买。今年9月智谱、阿里云百炼、千问办公等已开设旗舰店,天猫同期上线"AI空间站"充值中心。但各厂商旗舰店销量并不理想,智谱最高一款产品仅售出200多份。
