此语言仍有 24 条内容或界面任务待翻译;未完成部分显示中文原文。

Insights & News

AI、算力、具身智能机器人与智能硬件的产业动态,以及选型与采购指南。

今日更新0 篇Updates This Week23 篇Published147 篇订阅周报

2026-10-30 · 本站更新

具身智能机器人Industry Updates

Galaxy General and Unitree Technology Announce Jointly Developed Humanoid Robot Galaxy Xingzai (Galbot ET1)

At the 2026 World Physical AI Conference and Unitree Partner Ecosystem Conference, Galaxy General Robotics announced the official launch of Galaxy Xingzai (Galbot ET1), a native embodied agent robot jointly developed with Unitree Technology. The product is powered by Galaxy General's proprietary AstraBrain-WBC cerebellum foundation model and can observe and learn human movements in real time; it is offered at three price points of 7.9 thousand, 13.9 thousand, and 17.9 thousand yuan, with reservations exceeding 200 units within 24 hours of opening. Galaxy General is responsible for product definition and embodied intelligence R&D, while Unitree provides joint actuator modules and hardware mass production support.

纺织服装机械网(转载“银河通用Galbot”微信公众号) · 约 7 分钟

大模型与软件Industry Updates

iFLYTEK Launches Xingchen TaaS Platform and Spark X2.6, with 11870K Developers

At the 2026 Global 1024 Developer Festival, iFLYTEK announced the device-cloud integrated Spark X2.6, the Xingchen TaaS platform, the SparkAudio 1.0 speech foundation model, and Agent OS. Spark X2.6 expands its context window to one million tokens and open-sources the 4B on-device model, with device-cloud routing accuracy improved to 96.1%; the iFLYTEK Open Platform developer base has reached 1187 ten thousand.

央广网 · 约 7 分钟

大模型与软件Industry Updates

Tencent Announces Hunyuan Hy4 Is Coming Soon with First-Time Visual Understanding

At the 2026 Tencent Global Digital Ecosystem Summit, Tencent announced that Hunyuan Hy4 is coming soon. Compared with the Hy4 preview, Hy4 reportedly delivers comprehensive, substantial improvements, with enhanced capabilities for long-horizon and real-world productivity tasks, improved chain-of-thought efficiency, and first-time visual understanding; the more practical Hy4 flash is also planned for near-term release. The article reviews the officially disclosed information and points still to be confirmed.

证券时报网(人民财讯) · 约 5 分钟

大模型与软件Industry Updates

Modelbest Releases Open-Source On-Device Multimodal Model MiniCPM-V 4.7, 35B-A3B MoE Runs on About 3 GB of Memory

On 10 13, Modelbest released the open-source on-device multimodal large model MiniCPM-V 4.7, available in 35B-A3B and 1B versions. 35B-A3B adopts a MoE architecture with 350 billion total parameters and 30 billion activated per inference; according to the company, it requires only about 3 GB of memory to run, with a decoding speed of 15+ token/s. It supported Rockchip, NVIDIA, Intel, and Apple chip platforms as well as iOS, Android, and HarmonyOS from day one, with core optimizations from the proprietary LES lookahead expert scheduling technology.

上海证券报·中国证券网 · 约 5 分钟

大模型与软件Industry Updates

DeepSeek V4.1 Flash Updated to Version V4.1-Flash-1017 with API Search and Fetch Interfaces Opened

According to the official news brief from DeepSeek, V4.1 Flash has been updated to version V4.1-Flash-1017, with the web version, APP, and API model simultaneously updated. According to the brief, the new model shows improved benchmark results while shortening task runtime and output length, lowering usage costs, and was optimized for platform frameworks such as DeepSeek Harness, Tencent CodeBuddy, and WorkBuddy. Through self-feedback reinforcement learning training, the model has enhanced its project-building capabilities in areas such as game development and 3D modeling; DeepSeek API adds support for the Search and Fetch interfaces, which are now officially open.

CFM 闪存市场 · 约 4 分钟

2026-10-29 · 本站更新

大模型与软件Industry Updates

Meta releases 8160-billion-parameter open-source model Muse Spark Open, says two Mac Studios can run it locally

According to BT Finance, Meta released the native multimodal open-source model Muse Spark Open on 2026/10/15, with 8160 billion total parameters and 420 billion active parameters. Built on a MoE architecture and deeply integrated with Apple's MLX framework, the model can reportedly run fully locally on two Mac Studios equipped with M5 Ultra and 512GB of memory. The article also notes that its performance data lacks independent third-party verification, and that on coding tasks it reportedly still lags some Chinese open-source models.

同一事件另有 1 篇报道

BT财经 · 约 8 分钟

大模型与软件Industry Updates

DeepSeek releases V4.1-Flash-1017: small-task cost drops to 40% of flagship, time doubles

On 2026 17, 10, DeepSeek upgraded V4.1-Flash-1017 in place, with the API name and pricing unchanged, focusing on enhanced agent capabilities and wish-style programming. Third-party tests show that when all 10 coding tasks passed, spending was under 40% of Claude Opus 5.5, but took more than twice as long; cache gains depend on multi-turn session reuse, benchmarks are tied to evaluation frameworks, and hallucination correction, latency, and low-price competition remain open gaps.

同一事件另有 1 篇报道

DOIT · 约 10 分钟

具身智能机器人Industry Updates

FF Launches 2490-Dollar Entry-Level Humanoid Robot FF Nova and Educational Nova Mini

On 10/16, Faraday Future and FF EAI Robotics Ecosystem jointly launched two humanoid robots: the 2490-dollar entry-level embodied AI robot FF Nova and the education-focused FF Nova Mini (base unit 299 dollars, courses 499 dollars). FF Nova has 20 degrees of freedom across its body, a multimodal sensor matrix, and an open ROS 2 SDK. The two parties also announced FFR's all-stock acquisition of FFAI's robotics business, expected to close by the end of the month.

机器人大讲堂 · 约 7 分钟

具身智能机器人Industry Updates

In 2026, embodied intelligence splits along two tracks: robots enter factories and homes alike, as Qiyuan enters with in-stock units

As Tech Stack noted in its 2026 month 10, 17 article, embodied intelligence has split into two routes—factory and home: first-half global humanoid shipments reached about 1.91,0 units with Chinese vendors over 97% share; consumer prices have fallen to the 10,000-yuan level but with clear spec trade-offs. Qiyuan Q1 and T1 went on sale from stock at 19999 yuan, with the line producing one unit every 2.5 minutes, and housework is not yet the first selling point. The article argues both routes face the shared challenge of environmental generalization.

同一事件另有 5 篇报道

技术栈 · 约 10 分钟

Compute and ServersIndustry Updates

Google Cloud 与 NVIDIA 宣布 Vera Rubin 基础设施正式可用

NVIDIA 博客宣布 Google Cloud 搭载 NVIDIA Vera Rubin NVL72 系统的 A5X 实例推出,正式可用定于下月,并计划扩展至欧洲。A5X 集成 ConnectX-9 SuperNIC、Google Cloud Virgo 网络与第四代液冷,面向大规模智能体推理与强化学习。Google Cloud 计划将 Vera Rubin 基础设施扩展至最多 960,000 块 Rubin GPU,Ineffable Intelligence 为首个客户。

NVIDIA Blog · 约 3 分钟

2026-10-28 · 本站更新

工业 AIIndustry Updates

SAMR Releases First Three National Standardization Guiding Technical Documents on AI Empowerment of the Steel Industry

SAMR has released three national standardization guiding technical documents — *Artificial Intelligence — Classification Guide for Application Scenarios in the Steel Industry*, *Artificial Intelligence — Technical Requirements for Steel Industry Large Model Datasets*, and *Artificial Intelligence — Service Capability Maturity Assessment for Steel Industry Large Models* — covering scenario classification, dataset specifications, and maturity assessment, building a complete closed loop for steel industry large models, supporting the digital-intelligent transformation of steel enterprises, and moving applications from pilots toward large-scale deployment.

现代物流产业网(现代物流报官网,转载央视新闻客户端) · 约 6 分钟

大模型与软件Industry Updates

Alibaba Cloud Bailian Platform Issues DeepSeek-V4.1-Flash Upgrade Notice

The announcements section of Alibaba Cloud's official website published an upgrade notice for "Model Studio (Bailian)" regarding DeepSeek-V4.1-Flash, dated 2026-10-21, with the affected product listed as Model Studio (Bailian). The current announcement page does not provide details on the upgrade content, model behavior changes, or effective timing; verification should be based on Bailian platform documentation or subsequent announcements.

阿里云 · 约 3 分钟

Product Selection Assistant

让需求变成可修改的筛选条件

依据已录入参数生成条件,结果可查看和修改。帧率与性能需核对实测,不能仅凭算力数值保证。