每经AI快讯，2月13日，淡水河谷2025年四季度净运营收入110.6亿美元，预期111.2亿美元；四季度净亏 损38.4亿美元，预计净利润26.1亿美元。 ...

每经AI快讯，2月13日，港交所文件显示，深圳易慕峰生物科技股份有限公司向港交所提交上市申请， 独家保荐人华泰国际。 ...

每经记者｜王琳 每经编辑｜金冥羽 胡玲 王晓波 标题点睛 美股三大指数集体收跌，道指跌1.34%，纳指跌2.03%，标普500指数跌1.57%，大型科技股普跌，苹果跌5%，博通跌超3%，亚马逊、Meta、特斯拉跌超 2%；存储概念股、食品百货涨幅居前，闪迪、希捷科技涨超5%，西部数据、沃尔玛涨超3%；贵金属、加密货币板块跌幅居前，美国黄金公司、泛美白 银、Robinhood跌超8%，Coinbase跌超7%，Circle跌超2%。中概股普遍下跌，纳斯达克中国金龙指数跌2.97%，腾讯音乐跌超10%，贝壳跌近6%，哔哩哔 哩跌近5%，百度集团跌超4%，世纪互联涨超3%，金山云涨逾2%。 国际贵金属价格大幅下挫，现货黄金一度跌破4900美元关口，截至发稿时，现货黄金跌3.27%，报4917.09美元/盎司；COMEX黄金期货跌3.19%，报4936 美元/盎司。现货白银跌10.84%，报75.07美元/盎司；COMEX白银期货跌10.93%，报74.75美元/盎司。 2月12日，外交部发言人林剑主持例行记者会。在回答有关安世半导体的问题时，林剑表示，中方主管部门已多次就安世半导体问题阐明立场，指出安世 半导体问题 ...

每经AI快讯，当地时间2月12日，洲际交易所（ICE）农产品期货主力合约收盘多数上涨，原糖期货涨 0.07%报13.53美分/磅，棉花期货涨0.39%报64.29美分/磅，可可期货跌4.07%报3699.00美元/吨，咖啡期 货涨2.12%报299.05美分/磅。 ...

每经AI快讯，2月13日，港股期指夜盘收盘，恒生指数期货跌1.05%，恒生科技指数期货跌1.46%。 ...

每经AI快讯，2月13日，上期所基本金属期货夜盘收盘全线下跌，沪锡跌4.27%，沪镍跌3.74%，沪铜跌 2.58%，不锈钢跌1.92%，沪铝跌0.91%，沪锌跌0.63%，氧化铝跌0.46%，沪铅跌0.30%。 ...

每经AI快讯，当地时间2月12日，芝加哥期货交易所（CBOT）农产品期货主力合约收盘全线上涨，大 豆期货涨1.11%报1136.50美分/蒲式耳，玉米期货涨0.88%报431.25美分/蒲式耳，小麦期货涨2.75%报 552.00美分/蒲式耳。 ...

Group 1 - The core viewpoint of the article indicates that all base metals in London experienced a decline on February 12, with specific percentage drops noted for each metal [1] Group 2 - LME aluminum fell by 0.18% to $3,097.5 per ton [1] - LME lead decreased by 0.45% to $1,984.0 per ton [1] - LME zinc dropped by 0.73% to $3,381.5 per ton [1] - LME copper saw a decline of 2.36% to $12,855.5 per ton [1] - LME nickel experienced a significant drop of 3.52% to $17,250.0 per ton [1] - LME tin fell by 3.70% to $47,800.0 per ton [1]

临近春节，车厘子异常热销。不过，近期有多名消费者反映，购买到的车厘子出现不同程度坏果。 2025年11月，来自南半球的新一季车厘子刚登陆中国市场时，山姆会员店"3J级"智利车厘子售价159.9元/公斤，到了2026年1月，同款产品价格已"腰斩"至 79.9元/公斤。在盒马门店，降价趋势更为显著：今年1月促销期间，"2J级"车厘子礼盒叠加返券后售99元/5斤（约合20元/斤），而前一个月同等规格产品 售价还在180元/5斤左右。 从每年11月到次年2月，随着车厘子供应量逐渐上升，价格随之下降，这符合价格随供需关系变化的规律。但拉长时间看，本产季的价格远低于往年。记 者从广州江南果菜批发市场获悉，2月5日的智利进口车厘子批发价同比去年下降约22%，较前年更是下跌39%，这一产季（2025年11月至2026年2月）的 价格创近五年新低。 然而，近期多位网友通过视频反映，今年购买的车厘子品质普遍不佳，口感发苦，虽然外表看起来新鲜，但内里已经变质，消费体验犹如"开盲盒"。在引 发关注的测评视频中，网友在多家超市购买了车厘子。开箱后发现，果实均出现了果肉褐变、口感变质的情况。 2月12日下午，极目新闻记者多方了解到，胖东 ...

Core Insights - DeepSeek has initiated a gray testing phase for its flagship model, allowing for a context length of up to 1 million tokens, significantly expanding from the previous 128K tokens in version V3.1 released in August last year [1] - Users have reported mixed reactions to the recent updates, with some expressing dissatisfaction over the model's change in tone and interaction style, leading to a trending topic on social media regarding its perceived coldness [1][4] Group 1: Model Updates and Features - The latest version of DeepSeek supports the processing of extremely long texts, as demonstrated by its ability to handle a document with over 240,000 tokens [1] - The upcoming DeepSeek V4 model is expected to be released in mid-February 2026, with the current version being a speed-optimized variant that sacrifices some quality for performance testing [6] - DeepSeek's V series models are designed for optimal performance, with V3 marking a significant milestone due to its efficient MoE architecture [6] Group 2: User Feedback and Reactions - Users have criticized the new version for its impersonal approach, referring to users as "users" instead of personalized nicknames, which has led to a perception of the model being less engaging [4] - Some users have described the updated model as overly simplistic and lacking emotional depth, comparing its output unfavorably to older literary styles [4] - Conversely, a segment of users appreciates the model's newfound objectivity and rationality, noting that it appears more attuned to the psychological state of the questioner [5] Group 3: Technical Innovations - DeepSeek has introduced two innovative architectures: mHC for optimizing information flow in deep Transformers, enhancing stability and scalability without increasing computational load, and Engram for decoupling static knowledge from dynamic computation [7] - These innovations aim to significantly reduce the cost of long-context reasoning while maintaining performance [7]