小米开源轻量级文档理解模型 Xiaomi-OCR-0，参数规模 0.8B，权重与推理代码已上线 Hugging Face 和 GitHub。该模型基于 Qwen3.5-0.8B-Base 构建，用约 1.7 亿条 OCR 与文档理解样本训练，单一模型覆盖文档解析、关键信息抽取与 OCR 视觉问答三类任务，原生支持 Markdown、LaTeX 与 JSON 输出。官方评测显示其在 OmniDocBench v1.6 得 96.83 分，位列总榜第二、1B 以下参数模型第一，同时提供浏览器 Demo、MCP 服务与 Agent Skill 工具包。
#模型发布#开源生态#多模态
#模型发布#开源生态#推理
法国 Mistral AI 发布 Large 4 模型（ML4），具备先进编程和网络防御能力，企业可下载开放权重版本并部署在自有基础设施上运行。该模型总参数量 1.05 万亿，每次执行任务实际激活约 520 亿参数，在欧洲数据中心使用 3800 颗英伟达 Grace Blackwell GPU 训练。Mistral 称其在编程任务上可与其他领先开放模型相媲美，并强调开源有助于避免前沿 AI 技术集中在少数参与者手中。
#模型发布#开源生态#Agent
智元发布新一代机器人具身模型GO-3 Preview，打通训练、仿真练习到上岗后持续学习的全流程，在RoboDojo、RoboTwin、RoboColiseum三项测试中排名第一，计划2026年底全面开源。在便利店真实场景测试中，精灵G2机器人任务首日成功率28.6%，五天后提升到85.7%；挂小挂钩任务的成功率比传统训练方式高17.7个百分点。
#模型发布#具身智能#开源生态
大晓机器人联合南洋理工大学 S-Lab 发布面向具体环境构建的世界模型 Mundo，可将 1 至 12 张图像拓展为可探索、可交互的三维场景。该模型把物体质量、摩擦等物理属性写入场景程序，由物理引擎计算运动与碰撞，并支持用代码编辑物体、属性和动作。Mundo 已用于酒店楼层与洗衣区域重建，为机器人导航和操作训练提供仿真环境。
#模型发布#具身智能#多模态
智元发布新一代具身基座模型 GO-3 Preview，采用自研 WRAM 架构，将任务理解、未来状态预测与动作生成整合到统一模型中，并首次提出衡量机器人实际工作产出的 Worken 评价框架。智元称该模型在 RoboDojo、RoboTwin 和 RoboColiseum 三项具身智能评测中综合排名第一。Worken 等于活跃机器人数量、自主工作时长与产出效能比三者的乘积；同期智元还发布 AHA3 硬件平台和灵渠 OS 1.0。
#模型发布#具身智能#数据/训练
Mistral AI 于 10 月 29 日将 Mistral Large 4（代号 Le Chonk）的开放权重发布到 Hugging Face，此前 10 月 6 日仅提供付费 API 访问。该模型总参数 1000 亿级规模中的 1 万亿参数、采用 MoE 架构，每个 token 激活 490 亿参数（含输入输出层为 520 亿），共 128 个专家、每 token 启用 4 个加 1 个共享专家。三周的强化学习后，其 Terminal-Bench 4.0 得分从 28% 升至 41%，在 Artificial Analysis 代码指数上为 48%，与 GLM-5.3 持平且每任务平均少用 47% token；API 价格降为每百万 token 输入 0.70 美元、输出 2.10 美元，新许可证允许商用但转售模型服务需另签协议。
#模型发布#开源生态#Hugging Face
开发者 Divyansh Agrawal 于 10 月 22 日发布实验性模型 Qwenputer，基于阿里 Qwen3.5-4B 构建，在模型内部实现 32 位 RISC-V CPU 与内存功能，无需调用外部工具即可执行程序和精确计算。测试中大整数乘法准确率从 55% 提升至 97.5%，长整数运算基准 NUPA 达到 100%，标准版仅 48.6%。代价是模型体积增至约 18.8GB、每条指令需生成多个词元导致耗时数秒，且需专用驱动，项目已在 Hugging Face 和 GitHub 开源。
#模型发布#开源生态#推理
百度于本月 27 日正式发布并开源文心 Unlimited OCR 2 文档解析模型，将文档解析上限从数十页推向“无限”，可一次性解析累计超过 3,600 页的《哈利·波特》全系列 7 部作品。该模型原生支持版面要素切分识别，具备跨页文案合并与表格识别合并能力，在 MPDocBench 上得 84.6 分、MarathonDocBench 上得 77.1 分，均实现端到端性能 SOTA。
#模型发布#开源生态#多模态
Edge0 发布开源权重语音克隆 TTS 模型 Audio8-TTS-Realistic，601M 参数，支持 23 种语言，权重与代码以 Apache 2.0 发布。模型只需 1 段参考音频及对应文本即可零样本克隆音色，也可在无参考音色时直接生成语音。自报结果显示其在 CV3 九语言上平均错误率 4.86%，低于 Qwen3-TTS-1.7B 的 7.73%，但在 MiniMax 基准上落后于 4B 级模型，泰语 WER 达 48.57%。
#模型发布#语音#开源生态
#模型发布#Google#评测/基准
1/ Excited to announce our new company Particle This is our first step in a series of models focused on developers. With particle-1, we focused on token efficiency and managed to keep parity in performance while using ~45% fewer tokens. Live now: https://t.co/m5MvOwR9yE https://t.co/qAM7lFUPvG
#大佬观点#行业动态#模型发布
I did say we'd be releasing our models periodically...introducing Gemini 4 Carbon, our next frontier model. Following Argon's release just two weeks ago, Carbon shows the pace at which we’re pushing the frontier, it ranks #1 on the recently updated v5.0 of @ArtificialAnlys Intelligence Index at 59. It also leads across key cybersecurity evaluations scoring 88.2% on SEC-bench Pro for finding vulnerabilities, 54.8% on ExploitGym for exploitation, and 71.0% on CWE-bench for patching. We’re seeing significant improvements in long-horizon tasks even compared to Argon, and we’ve strengthened frontier safeguards. As we did with Argon, we’re going to roll it out gradually + responsibly. Some trusted cyber defenders will get access starting this week. We'll be expanding it to more partners through our Fairwind Program, ahead of wider availability soon, so hang tight!
#其他#模型发布#Google
Gemini 4 Carbon 发布，在 Artificial Analysis v5 上排名第一。发布者称其作为日常主力模型使用体验很好，并表示期待它正式上线后用户能做出什么。
#模型发布#Google#评测/基准
Introducing Gemini 4 Carbon, it's a big jump over Argon which we released just 2 weeks ago. Gemini 4 Carbon is SOTA on cyber, long horizon tasks, ranks #1 on Artificial Analysis V5, significantly improves on usability in domains like coding, and is more aligned. https://t.co/Ikk9YedjCQ
#模型发布#Google#评测/基准
Gemini 4 Carbon is our next frontier model. It’s state-of-the-art across cybersecurity and long-horizon capabilities.
#模型发布#安全/对齐#Agent
Introducing Phonic-1, the most advanced speech-to-speech model for agentic tasks — from @phonic_co, the speech-to-speech model company. We’ve had a breakout year and raised a $27.5M Series A led by @firstround with participation from @Lux_Capital, @felicis, and Definition. https://t.co/syYLXbQ4lL
#模型发布#语音#Agent
Meet NVIDIA Nemotron 3.5 Transcribe 🎙️ An open-weight, promptable 30B MoE model with ~4B active parameters that brings transcription in 40 languages, speaker attribution, translation, contextual prompting and audio understanding into one checkpoint. Watch it keep track of who said what when everyone’s talking at once. 1/ 🧵
#模型发布#语音#开源生态
Today, we release the weights of a new and improved Le Chonk, advancing the open frontier for agentic coding and cybersecurity. 💻 Built to build: Le Chonk ranks #1 on @cognition’s FrontierCode among open models, and sets a new benchmark for Pareto efficiency on @artificialanalysis. 🛡 Built to defend: State-of-the-art in cyberdefense, put through the paces by the world's leading cyber vendors, including @CrowdStrike, @cloudflare, @rubrik, @zscaler, and others. Built for sovereignty: Make it yours. Download the weights on @huggingface , customize it with Mistral Forge, and deploy it wherever you need it. Read more from our team and customers 👇:
#模型发布#开源生态#推理
I did say we'd be releasing our models periodically...introducing Gemini 4 Carbon, our next frontier model. Following Argon's release just two weeks ago, Carbon shows the pace at which we’re pushing the frontier, it ranks #1 on the recently updated v5.0 of @ArtificialAnlys Intelligence Index at 59. It also leads across key cybersecurity evaluations scoring 88.2% on SEC-bench Pro for finding vulnerabilities, 54.8% on ExploitGym for exploitation, and 71.0% on CWE-bench for patching. We’re seeing significant improvements in long-horizon tasks even compared to Argon, and we’ve strengthened frontier safeguards. As we did with Argon, we’re going to roll it out gradually + responsibly. Some trusted cyber defenders will get access starting this week. We'll be expanding it to more partners through our Fairwind Program, ahead of wider availability soon, so hang tight!
#模型发布#Google#安全/对齐