2026北京亦庄人形机器人半程马拉松正式开赛 | DeepSeek-V4开源登场
英语万花筒 | 英语精讲
Robots Outrun Humans for the First Time: Beijing Half Marathon Signals a “Silicon Awakening”
机器人首次跑赢人类：北京半马见证“硅基觉醒”
The 2026 Beijing Yizhuang Humanoid Robot Half Marathon officially kicked off, setting a new record for participation and becoming the world’s first large-scale long-distance race for humanoid robots.
2026北京亦庄人形机器人半程马拉松正式开赛，参赛规模创下新高，成为全球首个大规模人形机器人长距离竞速赛事。
The “Lightning” robot from Team Qitian Dasheng finished the race in 50 minutes and 26 seconds, surpassing the human men’s half marathon world record of 57 minutes and 20 seconds.
来自齐天大圣队的“闪电”机器人以50分26秒完赛，成绩超过人类男子半马世界纪录57分20秒。
About 40% of the robots in this year’s race achieved fully autonomous navigation without human control, reflecting major advances in perception and algorithms.
本届赛事中，约四成机器人实现完全自主导航，摆脱人工遥控，体现出算法与感知系统的显著进步。
Compared with earlier tests where robots frequently fell, this year’s models showed clear improvements in balance, energy management, and path planning.
与早期测试中频繁跌倒相比，今年机器人在动态平衡、能源管理和路径规划上明显成熟。
The event is not just a competition but a milestone marking the transition of humanoid robots from laboratories to real-world environments.
这场比赛不仅是技术展示，更被视为人形机器人从实验室迈向真实复杂场景的重要节点。
DeepSeek-V4 Goes Open Source: One Million Tokens Redefine AI Limits
DeepSeek-V4开源登场：百万Token重塑大模型能力边界
Chinese AI company DeepSeek has released its new flagship model DeepSeek-V4 preview version and made it fully open source, drawing global attention.
国产AI公司深度求索（DeepSeek）发布新一代旗舰模型DeepSeek-V4预览版，并宣布全面开源，在全球AI社区引发震动。
The model natively supports a context window of one million tokens, equivalent to about 750,000 words, significantly expanding long-text understanding and reasoning capabilities.
该模型原生支持100万Token上下文，约相当于75万字输入长度，大幅拓展长文本理解与复杂推理能力。
It delivers major upgrades in code generation, logical reasoning, and agent capabilities, and is seen as comparable to top-tier closed-source models globally.
在代码生成、逻辑推理以及智能体（Agent）能力方面，V4实现跨越式提升，被认为可对标国际顶级闭源模型。
Industry experts believe long-context capability is becoming a key battleground in the next phase of AI competition.
业内普遍认为，长上下文能力正在成为下一阶段大模型竞争的关键指标。
This release strengthens the influence of the open-source camp and further intensifies global competition in AI.
此次发布进一步强化了开源阵营的技术影响力，也让全球AI竞争格局更加多元和激烈。
Breakthrough in Domestic Lithography: 28nm Enters Real Industrial Deployment
国产光刻机关键突破：28nm量产迈入真实产业链
Shanghai Micro Electronics has achieved mass production of 28nm immersion DUV lithography machines, marking a key breakthrough in domestic semiconductor equipment.
上海微电子实现28nm浸没式DUV光刻机量产，标志着国产半导体设备取得关键进展。
The machines have reached a yield rate of over 90%, with localization exceeding 85%, demonstrating stronger supply chain coordination.
该设备良率已超过90%，整体国产化率超过85%，显示产业链协同能力持续提升。
The first unit has entered the production line of 中芯国际 for real-world manufacturing validation.
首台设备已进入中芯国际产线，开始实际生产验证，迈出商业化关键一步。
Although 28nm is not the most advanced node, it remains widely used in automotive electronics and industrial control systems.
尽管28nm并非最先进制程，但在汽车电子、工业控制等领域仍具有广泛应用价值。
This progress indicates growing self-reliance in core semiconductor equipment and enhanced industrial resilience.
这一进展意味着中国在核心半导体设备领域正逐步增强自主能力，产业韧性显著提升。
New Secrets in Lunar Soil: Chang’e-5 Reveals Clues to Moon Evolution
月壤再现新发现：嫦娥五号揭开月球演化新线索
Scientists have discovered two previously unknown minerals—magnesium-rich and cerium-rich “Chang’e stones”—in lunar samples returned by 嫦娥五号.
科研人员在嫦娥五号带回的月球样品中，发现两种此前未知的新矿物——镁嫦娥石和铈嫦娥石。
This marks the first confirmed discovery of new mineral types on the Moon containing specific rare-earth element structures.
这是人类首次在月球上确认含有特定稀土元素结构的新型矿物类型。
Studies suggest these minerals formed during periods of lunar volcanic activity, offering key evidence for understanding the Moon’s geological evolution.
研究表明，这些矿物形成于月球火山活动阶段，为理解月球地质演化提供关键证据。
The findings also provide important references for assessing lunar resource potential, especially in rare-earth elements.
相关发现也为评估月球资源潜力提供重要参考，尤其是在稀土元素开发方面。
As future lunar missions advance, research on lunar samples continues to expand humanity’s understanding of extraterrestrial materials.
随着后续探月任务推进，月球样品研究正持续拓展人类对地外物质的认知边界。
