美股詳情

本頁面由Tiger Fintech (Singapore) Pte. Ltd.提供服務

Shift Technologies, Inc.

0.1703

0.0000

成交量：- -

成交額：3,939.25萬

市值：289.49萬

市盈率：-0.02

高：0.1703

開：0.1703

低：0.1703

收：0.1703

52周最高：7.13

52周最低：0.1000

股本：1,699.90萬

流通股本：1,258.65萬

量比：- -

換手率：- -

股息：- -

股息率：- -

每股收益(TTM):-10.4259

每股收益(LYR):-19.9167

淨資產收益率：-2221.90%

總資產收益率：-37.67%

市淨率：-0.02

市盈率(LYR):-0.01

資料載入中...

總覽公司新聞公告

啱啱，DeepSeek重要突破！大模型上下文緊箍咒打破

智譜推出 AutoGLM 2.0：手機 Agent 的「Manus 時刻」？

月之暗面又開源了！楊植麟合著提出新Agent框架，旗艦模型得分超GPT-4o

智譜發佈新一代開源視覺模型GLM-4.5V

华尔街见闻

智譜GLM-4.5完整技術報告：提出三個通用模型關鍵能力，公開12項測試成績

陳天橋聯手清華教授代季峯首發最強開源AI模型項目，全力打造下一個DeepSeek

站在DeepSeek肩膀上，小紅書開源首款多模態模型：看懂表情包與數學題，一手實測

DeepSeek V4借實習生獲獎論文「起飛」？梁文峯劍指上下文：處理速度提10倍、要「完美」準確率

硬核「吵」了30分鐘：這場大模型圓桌，把AI行業的分歧説透了

復旦聯合南洋理工提出基於視覺Grounding的多輪強化學習框架MGPO

AI打假AI，拿下SOTA丨廈大&騰訊優圖

任務級獎勵提升AppAgent思考力，淘天提出Mobile-R1，3B模型超32B

為什麼2025成了Agent落地元年？

OpenAI聯合創始人揭秘AI進化新方向：讓模型學會和人類一樣反思

Karpathy戳破強化學習神話，首提AI覆盤式進化！暴力試錯將死

豆蔻婦科大模型再突破：釘釘行業訓練平台+精標數據SFT ，準確率從 77.1%上升至 90.2%

vivo發端側多模態模型，只有3B可理解GUI界面，20項評測表現亮眼

智譜再融10億！獲上海國資押注，開源視覺大模型，能解説球賽，還會玩手機

10B級模型SOTA，超8倍參數「大」模型，智譜開源GLM-4.1V-Thinking

SuperCLUE推理榜驚現黑馬：原來中興是一家AI公司？

{"basename":"/hant","ssrTDKData":{"titleTemplate":"%s - 老虎證券","title":"老虎证券全球投资理财平台| 一站式投资美股新股港股A股","description":"老虎证券助您一站式投资美股，新股，港股，A股等全球金融理财产品。新加坡华人最信赖的在线投资平台，现在加入即享低费用，24/5 无时差炒美股投资理财！","keywords":"老虎證券,老虎證券開戶,老虎券商,老虎證券官網,老虎證券app,tigertrade老虎證券,股票,炒股,新加坡股票交易平臺,投資,投資理財","social":{"ogDescription":"老虎证券助您一站式投资美股，新股，港股，A股等全球金融理财产品。新加坡华人最信赖的在线投资平台，现在加入即享低费用，24/5 无时差炒美股投资理财！","ogImage":"https://c1.itigergrowtha.com/portal5/static/media/og-logo.be62fbe1.png","ogUrl":"https://www.itiger.com/hant/stock/SFT/news"},"companyName":"老虎證券"},"pageData":{"isMobile":false,"isTiger":false,"isTTM":true,"region":"SGP","license":"TBSG","edition":"fundamental","symbol":"SFT","isAnalysisPage":false},"__swrFallback__":{"@#url:\"https://hq.skytigris.cn/stock_info/detail/global\",params:#delay:false,,method:\"POST\",data:#items:@#symbol:\"SFT\",,,,,undefined,":{"symbol":"SFT","market":"US","secType":"STK","nameCN":"Shift Technologies, Inc.","latestPrice":0.1703,"timestamp":1697659200000,"preClose":0.1703,"halted":4,"volume":0,"delay":0,"floatShares":12586491,"shares":16999021,"eps":-10.42588,"marketStatus":"退市","change":0,"latestTime":"10-28 09:58:38 EDT","open":0.1703,"high":0.1703,"low":0.1703,"amount":39392485.363854,"amplitude":0,"askPrice":0,"askSize":0,"bidPrice":0,"bidSize":0,"shortable":3,"etf":0,"ttmEps":-10.42588,"tradingStatus":0,"nextMarketStatus":{"tag":"盤後交易","tradingStatus":3,"beginTime":1761681600000},"marketStatusCode":2,"adr":0,"exchange":"NASDAQ","adjPreClose":0.1703,"volumeRatio":0},"@#url:\"https://hq.skytigris.cn/stock_info/fundamental/all\",params:#delay:false,,method:\"POST\",data:#items:@#symbol:\"SFT\",,,,,undefined,":{"symbol":"SFT","floatShares":12586491,"roa":"-37.67%","roe":"-2221.90%","lyrEps":-19.91671,"shares":16999021,"dividePrice":0,"high":0.1703,"amplitude":0,"preClose":0.1703,"low":0.1703,"week52Low":0.1,"pbRate":"-0.02","psRate":"0.01","week52High":7.13,"institutionHeld":0,"latestPrice":0.1703,"eps":-10.42588,"divideRate":0,"volume":0,"delay":0,"ttmEps":-10.42588,"open":0.1703},"@#url:\"https://hq.skytigris.cn/fundamental/corporate_actions/details/SFT\",params:#limit:5,,,undefined,":[{"market":"US","date":"2023-08-10","symbol":"SFT","fiscalQuarterEnding":"2023/06","expectedEps":-1.85,"name":null,"time":"盤後","type":"earning","dateTimestamp":1691640000000,"reportTimeType":"post","actualEps":-1.44},{"market":"US","date":"2023-05-11","symbol":"SFT","fiscalQuarterEnding":"2023/03","expectedEps":-2.62,"name":null,"time":"盤後","type":"earning","dateTimestamp":1683777600000,"reportTimeType":"post","actualEps":-2.15},{"market":"US","date":"2023-03-28","symbol":"SFT","fiscalQuarterEnding":"2022/12","expectedEps":-5.07,"name":null,"time":"盤後","type":"earning","dateTimestamp":1679976000000,"reportTimeType":"post","actualEps":-3.65},{"market":"US","date":"2023-03-08","symbol":"SFT","defaultRemindTime":1678285800000,"type":"split","dateTimestamp":1678251600000,"forFactor":10,"toFactor":1,"ratio":10},{"market":"US","date":"2022-08-09","symbol":"SFT","fiscalQuarterEnding":"2022/06","expectedEps":-0.56,"name":null,"time":"盤後","type":"earning","dateTimestamp":1660017600000,"reportTimeType":"post","actualEps":-0.59}],"@#url:\"https://hq.skytigris.cn/fundamental/dividend/history\",params:#symbol:\"SFT\",market:\"US\",,,undefined,":[],"@#url:\"https://hq.skytigris.cn/fundamental/estimate/recommendation\",params:#symbol:\"SFT\",market:\"US\",delay:false,,,undefined,":{},"@#url:\"https://hq.skytigris.cn/value_analysis/stock/SFT\",params:#withQuantilePoint:true,period:\"5year\",factor:\"peRate\",delay:false,algorithm:1,,,undefined,":{"brief":{"id":"SFT","date":"2023-10-17","current":-0.016334,"percent":1,"low":-14.398732,"twenty":-8.041489,"median":-0.967796,"eighty":-0.156343,"high":-0.013458,"avg":-3.589974,"sd":4.206369,"marketCap":1839294},"quantilePoints":[{"date":"2020-10-30","current":-10.317712,"twenty":-13.113928,"median":-11.969705,"eighty":-10.922863,"marketCap":730751944},{"date":"2020-11-06","current":-10.07426,"twenty":-12.807875,"median":-11.071253,"eighty":-10.299163,"marketCap":713509482},{"date":"2020-11-13","current":-7.963277,"twenty":-12.596883,"median":-10.914143,"eighty":-10.27134,"marketCap":598559738},{"date":"2020-11-20","current":-7.198628,"twenty":-12.478635,"median":-10.526385,"eighty":-9.035106,"marketCap":541084866},{"date":"2020-11-27","current":-10.060601,"twenty":-12.061289,"median":-10.398862,"eighty":-7.976385,"marketCap":756205101},{"date":"2020-12-04","current":-9.776589,"twenty":-11.349483,"median":-10.27134,"eighty":-8.546596,"marketCap":734857292},{"date":"2020-12-11","current":-9.437958,"twenty":-11.09212,"median":-10.257225,"eighty":-9.398633,"marketCap":709404134},{"date":"2020-12-18","current":-9.437958,"twenty":-11.061979,"median":-10.067431,"eighty":-9.073111,"marketCap":709404134},{"date":"2020-12-24","current":-10.027831,"twenty":-11.04343,"median":-10.067431,"eighty":-9.285028,"marketCap":753741892},{"date":"2020-12-31","current":-9.033786,"twenty":-11.024881,"median":-10.055837,"eighty":-9.263181,"marketCap":679024559},{"date":"2021-01-08","current":-8.771621,"twenty":-10.803404,"median":-9.984136,"eighty":-8.935474,"marketCap":659318888},{"date":"2021-01-15","current":-9.077481,"twenty":-10.769836,"median":-9.787513,"eighty":-8.946398,"marketCap":682308837},{"date":"2021-01-22","current":-9.001016,"twenty":-10.575075,"median":-9.697582,"eighty":-8.944213,"marketCap":676561350},{"date":"2021-01-29","current":-8.957322,"twenty":-10.45219,"median":-9.645506,"eighty":-8.955137,"marketCap":673277072},{"date":"2021-02-05","current":-8.924551,"twenty":-10.405818,"median":-9.547194,"eighty":-8.944213,"marketCap":670813863},{"date":"2021-02-12","current":-11.087416,"twenty":-10.735518,"median":-9.645506,"eighty":-8.955137,"marketCap":833385644},{"date":"2021-02-19","current":-11.600824,"twenty":-10.996861,"median":-9.749658,"eighty":-8.983538,"marketCap":871975915},{"date":"2021-02-26","current":-9.667353,"twenty":-10.932302,"median":-9.787513,"eighty":-9.014124,"marketCap":726646596},{"date":"2021-03-05","current":-8.749774,"twenty":-10.842729,"median":-9.776589,"eighty":-8.983538,"marketCap":657676749},{"date":"2021-03-12","current":-14.398732,"twenty":-11.028048,"median":-9.814821,"eighty":-9.014124,"marketCap":851627427},{"date":"2021-03-19","current":-12.539463,"twenty":-11.094862,"median":-9.984136,"eighty":-9.022863,"marketCap":741659070},{"date":"2021-03-26","current":-12.027648,"twenty":-11.389178,"median":-10.055837,"eighty":-9.022863,"marketCap":711387271},{"date":"2021-04-01","current":-11.643787,"twenty":-11.507717,"median":-10.105663,"eighty":-9.027232,"marketCap":688683422},{"date":"2021-04-09","current":-12.411509,"twenty":-11.72909,"median":-10.169837,"eighty":-9.051264,"marketCap":734091120},{"date":"2021-04-16","current":-10.819196,"twenty":-11.73762,"median":-10.213531,"eighty":-9.077481,"marketCap":639912191},{"date":"2021-04-23","current":-10.597948,"twenty":-11.656583,"median":-10.191684,"eighty":-9.077481,"marketCap":626826237},{"date":"2021-04-30","current":-11.992041,"twenty":-11.88832,"median":-10.264283,"eighty":-9.094959,"marketCap":709281232},{"date":"2021-05-07","current":-11.366121,"twenty":-11.97785,"median":-10.30092,"eighty":-9.134283,"marketCap":672260622},{"date":"2021-05-14","current":-7.191308,"twenty":-11.953683,"median":-10.28613,"eighty":-9.077481,"marketCap":644125458},{"date":"2021-05-21","current":-7.017514,"twenty":-11.94087,"median":-10.257225,"eighty":-9.022863,"marketCap":628558742},{"date":"2021-05-28","current":-6.651138,"twenty":-11.88832,"median":-10.186222,"eighty":-8.950768,"marketCap":595742422},{"date":"2021-06-04","current":-7.111457,"twenty":-11.820079,"median":-10.174632,"eighty":-8.880857,"marketCap":636973183},{"date":"2021-06-11","current":-7.872392,"twenty":-11.761789,"median":-10.150994,"eighty":-8.77599,"marketCap":705130155},{"date":"2021-06-18","current":-7.750267,"twenty":-11.660848,"median":-10.067431,"eighty":-8.135607,"marketCap":694191382},{"date":"2021-06-25","current":-8.08846,"twenty":-11.635194,"median":-10.027831,"eighty":-7.997621,"marketCap":724483370},{"date":"2021-07-02","current":-7.919363,"twenty":-11.592085,"median":-9.887162,"eighty":-7.969831,"marketCap":709337376},{"date":"2021-07-09","current":-8.351499,"twenty":-11.557129,"median":-9.814821,"eighty":-7.947546,"marketCap":748043804},{"date":"2021-07-16","current":-7.750267,"twenty":-11.511784,"median":-9.776589,"eighty":-7.919363,"marketCap":694191382},{"date":"2021-07-23","current":-7.938152,"twenty":-11.501616,"median":-9.683739,"eighty":-7.909969,"marketCap":711020264},{"date":"2021-07-30","current":-7.938152,"twenty":-11.476428,"median":-9.645506,"eighty":-7.909969,"marketCap":711020264},{"date":"2021-08-06","current":-8.041489,"twenty":-11.458965,"median":-9.547194,"eighty":-7.909969,"marketCap":720276149},{"date":"2021-08-13","current":-6.282,"twenty":-11.373662,"median":-9.437958,"eighty":-7.862998,"marketCap":642259065},{"date":"2021-08-20","current":-5.779769,"twenty":-11.366121,"median":-9.399726,"eighty":-7.825421,"marketCap":590912010},{"date":"2021-08-27","current":-6.1338,"twenty":-11.331011,"median":-9.263181,"eighty":-7.769055,"marketCap":627107475},{"date":"2021-09-03","current":-6.644264,"twenty":-11.266543,"median":-9.214025,"eighty":-7.71269,"marketCap":679296285},{"date":"2021-09-10","current":-6.693664,"twenty":-11.263063,"median":-9.170331,"eighty":-7.657199,"marketCap":684346815},{"date":"2021-09-17","current":-6.520765,"twenty":-11.235977,"median":-9.077481,"eighty":-7.333643,"marketCap":666669960},{"date":"2021-09-24","current":-6.306699,"twenty":-11.168077,"median":-9.055633,"eighty":-7.204045,"marketCap":644784330},{"date":"2021-10-01","current":-5.63157,"twenty":-11.115373,"median":-9.022863,"eighty":-7.10582,"marketCap":575760420},{"date":"2021-10-08","current":-5.48337,"twenty":-11.091139,"median":-8.979169,"eighty":-6.97419,"marketCap":560608830},{"date":"2021-10-15","current":-5.60687,"twenty":-11.082774,"median":-8.935474,"eighty":-6.846103,"marketCap":573235155},{"date":"2021-10-22","current":-5.53277,"twenty":-11.073325,"median":-8.848086,"eighty":-6.640971,"marketCap":565659360},{"date":"2021-10-29","current":-5.623336,"twenty":-11.052704,"median":-8.793468,"eighty":-6.608037,"marketCap":574918665},{"date":"2021-11-05","current":-6.026768,"twenty":-11.035852,"median":-8.711541,"eighty":-6.486185,"marketCap":616164660},{"date":"2021-11-12","current":-4.263503,"twenty":-11.026464,"median":-8.426654,"eighty":-6.395619,"marketCap":495875283},{"date":"2021-11-19","current":-3.585536,"twenty":-10.987521,"median":-8.304528,"eighty":-6.301759,"marketCap":417022983},{"date":"2021-11-26","current":-3.313625,"twenty":-10.932302,"median":-8.262254,"eighty":-6.234246,"marketCap":385397787},{"date":"2021-12-03","current":-2.99904,"twenty":-10.897569,"median":-8.135431,"eighty":-6.13874,"marketCap":348809389},{"date":"2021-12-10","current":-2.761354,"twenty":-10.8665,"median":-8.08846,"eighty":-6.066287,"marketCap":321164822},{"date":"2021-12-17","current":-2.278991,"twenty":-10.83057,"median":-8.041489,"eighty":-6.031708,"marketCap":265062613},{"date":"2021-12-23","current":-2.418806,"twenty":-10.812879,"median":-8.041489,"eighty":-5.964195,"marketCap":281324123},{"date":"2021-12-31","current":-2.383853,"twenty":-10.789977,"median":-7.990585,"eighty":-5.876922,"marketCap":277258745},{"date":"2022-01-07","current":-2.125194,"twenty":-10.752677,"median":-7.966335,"eighty":-5.804469,"marketCap":247174952},{"date":"2022-01-14","current":-1.831582,"twenty":-10.7182,"median":-7.952243,"eighty":-5.689203,"marketCap":213025781},{"date":"2022-01-21","current":-1.391163,"twenty":-10.687614,"median":-7.938152,"eighty":-5.63157,"marketCap":161802024},{"date":"2022-01-28","current":-1.391163,"twenty":-10.613752,"median":-7.909969,"eighty":-5.610163,"marketCap":161802024},{"date":"2022-02-04","current":-1.412136,"twenty":-10.583635,"median":-7.900575,"eighty":-5.573937,"marketCap":164241251},{"date":"2022-02-11","current":-1.328246,"twenty":-10.508175,"median":-7.862998,"eighty":-5.48337,"marketCap":154484345},{"date":"2022-02-18","current":-1.251348,"twenty":-10.434996,"median":-7.834815,"eighty":-5.368104,"marketCap":145540514},{"date":"2022-02-25","current":-1.328246,"twenty":-10.433641,"median":-7.825421,"eighty":-4.172642,"marketCap":154484345},{"date":"2022-03-04","current":-1.174449,"twenty":-10.398862,"median":-7.778449,"eighty":-3.480696,"marketCap":136596684},{"date":"2022-03-11","current":-1.18144,"twenty":-10.344614,"median":-7.750267,"eighty":-3.222745,"marketCap":137409759},{"date":"2022-03-18","current":-1.317001,"twenty":-10.317712,"median":-7.71269,"eighty":-3.040985,"marketCap":218975116},{"date":"2022-03-25","current":-1.222217,"twenty":-10.30092,"median":-7.668959,"eighty":-2.852234,"marketCap":203215544},{"date":"2022-04-01","current":-1.047614,"twenty":-10.27134,"median":-7.64693,"eighty":-2.495705,"marketCap":174184752},{"date":"2022-04-08","current":-1.007705,"twenty":-10.27134,"median":-7.331348,"eighty":-2.411816,"marketCap":167549142},{"date":"2022-04-14","current":-0.922898,"twenty":-10.260048,"median":-7.225936,"eighty":-2.309751,"marketCap":153448472},{"date":"2022-04-22","current":-0.768251,"twenty":-10.239747,"median":-7.198628,"eighty":-2.255222,"marketCap":127735484},{"date":"2022-04-29","current":-0.667277,"twenty":-10.200423,"median":-7.134942,"eighty":-2.125194,"marketCap":110946814},{"date":"2022-05-06","current":-0.587602,"twenty":-10.182946,"median":-7.022996,"eighty":-1.997962,"marketCap":97699433},{"date":"2022-05-13","current":-0.563219,"twenty":-10.179695,"median":-6.965571,"eighty":-1.662405,"marketCap":101695320},{"date":"2022-05-20","current":-0.469349,"twenty":-10.171755,"median":-6.914177,"eighty":-1.50721,"marketCap":84746100},{"date":"2022-05-27","current":-0.502203,"twenty":-10.161098,"median":-6.728766,"eighty":-1.435904,"marketCap":90678327},{"date":"2022-06-03","current":-0.474042,"twenty":-10.154162,"median":-6.647701,"eighty":-1.412136,"marketCap":85593561},{"date":"2022-06-10","current":-0.359991,"twenty":-10.142637,"median":-6.627798,"eighty":-1.391163,"marketCap":65000258},{"date":"2022-06-17","current":-0.299163,"twenty":-10.099382,"median":-6.590747,"eighty":-1.353413,"marketCap":54017164},{"date":"2022-06-24","current":-0.427483,"twenty":-10.068796,"median":-6.512531,"eighty":-1.33173,"marketCap":77186747},{"date":"2022-07-01","current":-0.407207,"twenty":-10.05679,"median":-6.479598,"eighty":-1.315359,"marketCap":73525716},{"date":"2022-07-08","current":-0.457709,"twenty":-10.046425,"median":-6.430199,"eighty":-1.289098,"marketCap":82644396},{"date":"2022-07-15","current":-0.506897,"twenty":-10.019092,"median":-6.339633,"eighty":-1.228978,"marketCap":91525788},{"date":"2022-07-22","current":-0.478736,"twenty":-9.968843,"median":-6.306699,"eighty":-1.198218,"marketCap":86441022},{"date":"2022-07-29","current":-0.474042,"twenty":-9.899467,"median":-6.2581,"eighty":-1.151401,"marketCap":85593561},{"date":"2022-08-05","current":-0.657088,"twenty":-9.857378,"median":-6.209608,"eighty":-1.118523,"marketCap":118644540},{"date":"2022-08-12","current":-0.51882,"twenty":-9.818098,"median":-6.137917,"eighty":-1.07118,"marketCap":104334147},{"date":"2022-08-19","current":-0.413369,"twenty":-9.807174,"median":-6.125567,"eighty":-1.009701,"marketCap":83128019},{"date":"2022-08-26","current":-0.388651,"twenty":-9.794067,"median":-6.047351,"eighty":-0.958817,"marketCap":78157303},{"date":"2022-09-02","current":-0.344404,"twenty":-9.771203,"median":-6.035001,"eighty":-0.881992,"marketCap":69259212},{"date":"2022-09-09","current":-0.35769,"twenty":-9.749658,"median":-6.018534,"eighty":-0.778228,"marketCap":71931184},{"date":"2022-09-16","current":-0.31319,"twenty":-9.6892,"median":-5.940318,"eighty":-0.702135,"marketCap":62982198},{"date":"2022-09-23","current":-0.275185,"twenty":-9.678277,"median":-5.886802,"eighty":-0.667277,"marketCap":55339510},{"date":"2022-09-30","current":-0.295305,"twenty":-9.667353,"median":-5.837402,"eighty":-0.587602,"marketCap":59385639},{"date":"2022-10-07","current":-0.269533,"twenty":-9.645506,"median":-5.804469,"eighty":-0.553832,"marketCap":54202861},{"date":"2022-10-14","current":-0.233216,"twenty":-9.590888,"median":-5.713903,"eighty":-0.539751,"marketCap":46899471},{"date":"2022-10-21","current":-0.204575,"twenty":-9.590888,"median":-5.65627,"eighty":-0.523038,"marketCap":41139887},{"date":"2022-10-28","current":-0.224864,"twenty":-9.5035,"median":-5.63157,"eighty":-0.516284,"marketCap":45219946},{"date":"2022-11-04","current":-0.196644,"twenty":-9.481653,"median":-5.623336,"eighty":-0.506897,"marketCap":39544897},{"date":"2022-11-11","current":-0.137377,"twenty":-9.437958,"median":-5.598636,"eighty":-0.492816,"marketCap":32904556},{"date":"2022-11-18","current":-0.125515,"twenty":-9.437958,"median":-5.573937,"eighty":-0.488123,"marketCap":30063382},{"date":"2022-11-25","current":-0.104221,"twenty":-9.42922,"median":-5.516304,"eighty":-0.478736,"marketCap":24962962},{"date":"2022-12-02","current":-0.093609,"twenty":-9.383341,"median":-5.479254,"eighty":-0.474042,"marketCap":22421310},{"date":"2022-12-09","current":-0.093681,"twenty":-9.302506,"median":-5.401038,"eighty":-0.463715,"marketCap":22438425},{"date":"2022-12-16","current":-0.068313,"twenty":-9.26755,"median":-4.221567,"eighty":-0.445697,"marketCap":16362421},{"date":"2022-12-23","current":-0.057988,"twenty":-9.236965,"median":-3.969951,"eighty":-0.42845,"marketCap":13889231},{"date":"2022-12-30","current":-0.0532,"twenty":-9.230411,"median":-3.585536,"eighty":-0.426187,"marketCap":12742492},{"date":"2023-01-06","current":-0.081497,"twenty":-9.217303,"median":-3.404505,"eighty":-0.413051,"marketCap":19520232},{"date":"2023-01-13","current":-0.077889,"twenty":-9.19327,"median":-3.268185,"eighty":-0.404501,"marketCap":18655899},{"date":"2023-01-20","current":-0.066098,"twenty":-9.180162,"median":-3.149342,"eighty":-0.393142,"marketCap":15831841},{"date":"2023-01-27","current":-0.071172,"twenty":-9.147392,"median":-3.040985,"eighty":-0.372549,"marketCap":17047041},{"date":"2023-02-03","current":-0.097647,"twenty":-9.112436,"median":-2.957095,"eighty":-0.360585,"marketCap":23388336},{"date":"2023-02-10","current":-0.103435,"twenty":-9.077481,"median":-2.761354,"eighty":-0.354215,"marketCap":24774692},{"date":"2023-02-17","current":-0.091287,"twenty":-9.077481,"median":-2.4992,"eighty":-0.338644,"marketCap":21865056},{"date":"2023-02-24","current":-0.07503,"twenty":-9.077481,"median":-2.436284,"eighty":-0.334787,"marketCap":17971279},{"date":"2023-03-03","current":-0.07503,"twenty":-9.033786,"median":-2.411816,"eighty":-0.324705,"marketCap":17971279},{"date":"2023-03-10","current":-0.090144,"twenty":-9.022863,"median":-2.345403,"eighty":-0.307424,"marketCap":21591319},{"date":"2023-03-17","current":-0.097242,"twenty":-9.022863,"median":-2.285982,"eighty":-0.291045,"marketCap":23291423},{"date":"2023-03-24","current":-0.096532,"twenty":-9.022863,"median":-2.230055,"eighty":-0.287376,"marketCap":23121413},{"date":"2023-03-31","current":-0.112052,"twenty":-9.001016,"median":-2.153157,"eighty":-0.283284,"marketCap":19277585},{"date":"2023-04-06","current":-0.114053,"twenty":-8.9748,"median":-2.125194,"eighty":-0.275868,"marketCap":19621828},{"date":"2023-04-14","current":-0.138195,"twenty":-8.950768,"median":-2.041305,"eighty":-0.240969,"marketCap":23775301},{"date":"2023-04-21","current":-0.144203,"twenty":-8.942028,"median":-1.90848,"eighty":-0.232507,"marketCap":24809009},{"date":"2023-04-28","current":-0.144203,"twenty":-8.933289,"median":-1.747693,"eighty":-0.223818,"marketCap":24809009},{"date":"2023-05-05","current":-0.159224,"twenty":-8.926736,"median":-1.660308,"eighty":-0.221768,"marketCap":27393281},{"date":"2023-05-12","current":-0.18426,"twenty":-8.880857,"median":-1.530979,"eighty":-0.212474,"marketCap":31700401},{"date":"2023-05-19","current":-0.140497,"twenty":-8.834978,"median":-1.485539,"eighty":-0.204117,"marketCap":22913877},{"date":"2023-05-26","current":-0.163236,"twenty":-8.813131,"median":-1.44709,"eighty":-0.197028,"marketCap":26622384},{"date":"2023-06-02","current":-0.160117,"twenty":-8.800022,"median":-1.415631,"eighty":-0.184832,"marketCap":26113676},{"date":"2023-06-09","current":-0.197547,"twenty":-8.784729,"median":-1.412136,"eighty":-0.17301,"marketCap":32218172},{"date":"2023-06-16","current":-0.233937,"twenty":-8.762882,"median":-1.394658,"eighty":-0.180691,"marketCap":38153099},{"date":"2023-06-23","current":-0.220421,"twenty":-8.734481,"median":-1.384173,"eighty":-0.18426,"marketCap":35948697},{"date":"2023-06-30","current":-0.22354,"twenty":-8.673309,"median":-1.359054,"eighty":-0.18426,"marketCap":36457405},{"date":"2023-07-07","current":-0.198587,"twenty":-8.66457,"median":-1.345581,"eighty":-0.185403,"marketCap":32387742},{"date":"2023-07-14","current":-0.251613,"twenty":-8.4738,"median":-1.328246,"eighty":-0.18769,"marketCap":41035777},{"date":"2023-07-21","current":-0.22562,"twenty":-8.36653,"median":-1.317001,"eighty":-0.193057,"marketCap":36796544},{"date":"2023-07-28","current":-0.18819,"twenty":-8.336469,"median":-1.314265,"eighty":-0.193057,"marketCap":30692048},{"date":"2023-08-04","current":-0.210024,"twenty":-8.287618,"median":-1.272321,"eighty":-0.193473,"marketCap":34253004},{"date":"2023-08-11","current":-0.201,"twenty":-8.26883,"median":-1.251348,"eighty":-0.189853,"marketCap":27470231},{"date":"2023-08-18","current":-0.154234,"twenty":-8.259436,"median":-1.219723,"eighty":-0.185403,"marketCap":21078786},{"date":"2023-08-25","current":-0.148014,"twenty":-8.159856,"median":-1.195422,"eighty":-0.180258,"marketCap":20228834},{"date":"2023-09-01","current":-0.149258,"twenty":-8.135431,"median":-1.170954,"eighty":-0.169267,"marketCap":20398825},{"date":"2023-09-08","current":-0.154234,"twenty":-8.112885,"median":-1.144893,"eighty":-0.1672,"marketCap":21078786},{"date":"2023-09-15","current":-0.157965,"twenty":-8.092218,"median":-1.118523,"eighty":-0.16282,"marketCap":21588756},{"date":"2023-09-22","current":-0.143039,"twenty":-8.08846,"median":-1.102489,"eighty":-0.15976,"marketCap":19548874},{"date":"2023-09-29","current":-0.169159,"twenty":-8.07155,"median":-1.069588,"eighty":-0.158632,"marketCap":23118668},{"date":"2023-10-06","current":-0.220156,"twenty":-8.049004,"median":-1.032649,"eighty":-0.159156,"marketCap":30088267},{"date":"2023-10-13","current":-0.016667,"twenty":-8.041489,"median":-0.977773,"eighty":-0.157578,"marketCap":2277868},{"date":"2023-10-17","current":-0.013458,"twenty":-8.041489,"median":-0.967796,"eighty":-0.156343,"marketCap":1839294}],"updateTime":1761659716268},"$inf$@#url:\"https://stock-news.skytigris.cn/v2/news/list\",params:#symbols:\"SFT\",pageSize:20,pageCount:1,lang_content:\"cn\",edition:\"fundamental\",,,undefined,":[{"items":[{"id":"2577843120","title":"啱啱，DeepSeek重要突破！大模型上下文緊箍咒打破","url":"https://stock-news.laohu8.com/highlight/detail?id=2577843120","media":"智东西","labels":[],"top":-1,"itemType":null,"share":"https://ttm.financial/m/news/2577843120?lang=zh_tw&edition=fundamental","pubTime":"2025-10-21 11:06","pubTimestamp":1761015970,"startTime":"0","endTime":"0","summary":"目前，这一模型已在Hugging Face上开源，而介绍DeepSeek-OCR模型技术细节与背后理论的技术报告也已同步公开。DeepSeek-OCR团队称，他们此番开源的模型是对一种潜在解决方案的初步探索，即利用视觉模态作为文本信息的高效压缩媒介。DeepSeek-OCR的架构分为两部分。当DeepEncoder训练完成后，DeepSeek-OCR团队使用多模态数据和纯文本数据，采用流水线并行策略来训练完整的模型。实验结果显示，在10×压缩率下，DeepSeek-OCR的解码精度可达约97%。","market":"hk","thumbnail":null,"type":0,"news_type":0,"thumbnails":[],"rights":null,"property":[],"language":"zh","translate_title":"","themeId":null,"theme_name":"","theme_type":"","isJumpTheme":false,"source_url":"https://tech.ifeng.com/c/8nd8NmIG6UZ","is_publish_highlight":false,"source_rank":0,"column":"","sentiment":"0","news_top_title":null,"news_tag":"","news_rank":0,"length":0,"strategy_id":0,"source":"fenghuang_stock","symbols":["BK4214","SFT"],"gpt_icon":0},{"id":"2560295411","title":"智譜推出 AutoGLM 2.0：手機 Agent 的「Manus 時刻」？","url":"https://stock-news.laohu8.com/highlight/detail?id=2560295411","media":"市场资讯","labels":[],"top":-1,"itemType":null,"share":"https://ttm.financial/m/news/2560295411?lang=zh_tw&edition=fundamental","pubTime":"2025-08-20 12:34","pubTimestamp":1755664440,"startTime":"0","endTime":"0","summary":"作者｜连冉编辑｜郑玄8 月 18 日，智谱正式发布了新的 ToC 产品 AutoGLM 2.0——一个手机通用 Agent。现在，当 AutoGLM 再次进入公众视野，情形已有所不同。AutoGLM 的新方案是用“云端原生”取代“本地镜像”。在与团队的交流中，极客公园了解到 AutoGLM 的产品哲学可以被提炼为“3A 原则”。这三大原则，共同构成了 AutoGLM 对一个成熟 Agent 形态的定义，也解释了其当前产品架构。据智谱披露，通过在线强化学习，AutoGLM 的任务成功率相较于冷启动阶段提升了 165%，超过 66% 的","market":"fut","thumbnail":null,"type":0,"news_type":0,"thumbnails":[],"rights":null,"property":[],"language":"zh","translate_title":"","themeId":null,"theme_name":"","theme_type":"","isJumpTheme":false,"source_url":"https://finance.sina.com.cn/stock/t/2025-08-20/doc-infmrakf1841118.shtml","is_publish_highlight":false,"source_rank":0,"column":"","sentiment":"0","news_top_title":null,"news_tag":"","news_rank":0,"length":0,"strategy_id":0,"source":"sina","symbols":["BK4214","SFT"],"gpt_icon":0},{"id":"2559312449","title":"月之暗面又開源了！楊植麟合著提出新Agent框架，旗艦模型得分超GPT-4o","url":"https://stock-news.laohu8.com/highlight/detail?id=2559312449","media":"智东西","labels":[],"top":-1,"itemType":null,"share":"https://ttm.financial/m/news/2559312449?lang=zh_tw&edition=fundamental","pubTime":"2025-08-14 13:10","pubTimestamp":1755148219,"startTime":"0","endTime":"0","summary":"在此基础上，研究人员提出了这一用于扩展CUA数据和基础模型的综合开源框架。其中旗舰模型OpenCUA-32B在CUA基准测试OSWorld-Verified上的平均成功率达到34.8%，达到开源模型新的SOTA，甚至超越了OpenAI CUA。OpenCUA-7B、OpenCUA-32B的平均分数均优于OpenAI、Qwen等模型。","market":"hk","thumbnail":null,"type":0,"news_type":0,"thumbnails":[],"rights":null,"property":[],"language":"zh","translate_title":"","themeId":null,"theme_name":"","theme_type":"","isJumpTheme":false,"source_url":"https://tech.ifeng.com/c/8loIqlUEjL7","is_publish_highlight":false,"source_rank":0,"column":"","sentiment":"0","news_top_title":null,"news_tag":"","news_rank":0,"length":0,"strategy_id":0,"source":"fenghuang_stock","symbols":["SFT","BK4214"],"gpt_icon":0},{"id":"2558666379","title":"智譜發佈新一代開源視覺模型GLM-4.5V","url":"https://stock-news.laohu8.com/highlight/detail?id=2558666379","media":"华尔街见闻","labels":[],"top":-1,"itemType":null,"share":"https://ttm.financial/m/news/2558666379?lang=zh_tw&edition=fundamental","pubTime":"2025-08-11 21:46","pubTimestamp":1754919965,"startTime":"0","endTime":"0","summary":"智谱表示推出全球 100B 级效果最佳的开源视觉推理模型 GLM-4.5V（总参数 106B，激活参数 12B），并同步在魔搭社区与 Hugging Face 开源。API 调用价格：低至输入 2 元/M tokens，输出 6 元/M tokens。 响应速度：达到 60-80 tokens/s。","market":"hk","thumbnail":"https://wpimg-wscn.awtmt.com/c1f578fd-c306-4f62-a0f1-fee307ada444.png","type":0,"news_type":0,"thumbnails":["https://wpimg-wscn.awtmt.com/c1f578fd-c306-4f62-a0f1-fee307ada444.png"],"rights":{"source":"wallstreetcn_hot_news","url":"https://wallstreetcn.com/articles/3753102","rn_cache_url":null,"directOrigin":true},"property":[],"language":"zh","translate_title":"","themeId":null,"theme_name":"","theme_type":"","isJumpTheme":false,"source_url":"https://wallstreetcn.com/articles/3753102","is_publish_highlight":false,"source_rank":0,"column":"","sentiment":"0","news_top_title":null,"news_tag":"","news_rank":0,"length":0,"strategy_id":0,"source":"wallstreetcn_hot_news","symbols":["LU0498741114.HKD","BK4214","SFT","BK4017","LU0006061336.USD","AGI","LU0368265764.SGD","LU0496367417.USD","BK4588","RL","BK4202","BK4585","LU0055631609.USD","LU0498741890.SGD"],"gpt_icon":0},{"id":"2558652406","title":"智譜GLM-4.5完整技術報告：提出三個通用模型關鍵能力，公開12項測試成績","url":"https://stock-news.laohu8.com/highlight/detail?id=2558652406","media":"智东西","labels":[],"top":-1,"itemType":null,"share":"https://ttm.financial/m/news/2558652406?lang=zh_tw&edition=fundamental","pubTime":"2025-08-11 18:07","pubTimestamp":1754906826,"startTime":"0","endTime":"0","summary":"编译 | 陈骏达编辑 | 云鹏智东西8月11日报道，近日，智谱发布了其最新一代旗舰模型GLM-4.5的完整技术报告。智东西此前已对GLM-4.5的能力进行了介绍与测试，在技术报告中，智谱进一步分享了这款模型在预训练、中期训练和后训练阶段进行的创新。值得一提的是，智谱还计划在今晚开源GLM-4.5系列的新模型，名为GLM-4.5V，或为一款视觉模型。SFT之后，GLM-4.5又进行了强化学习训练。","market":"hk","thumbnail":null,"type":0,"news_type":0,"thumbnails":[],"rights":null,"property":[],"language":"zh","translate_title":"","themeId":null,"theme_name":"","theme_type":"","isJumpTheme":false,"source_url":"https://tech.ifeng.com/c/8ljg4SI98PH","is_publish_highlight":false,"source_rank":0,"column":"","sentiment":"0","news_top_title":null,"news_tag":"","news_rank":0,"length":0,"strategy_id":0,"source":"fenghuang_stock","symbols":["LU0496367417.USD","LU0498741890.SGD","LU0368265764.SGD","BK4017","BK4214","LU0055631609.USD","SFT","AGI","LU0498741114.HKD"],"gpt_icon":0},{"id":"2558687160","title":"陳天橋聯手清華教授代季峯首發最強開源AI模型項目，全力打造下一個DeepSeek","url":"https://stock-news.laohu8.com/highlight/detail?id=2558687160","media":"市场资讯","labels":[],"top":-1,"itemType":null,"share":"https://ttm.financial/m/news/2558687160?lang=zh_tw&edition=fundamental","pubTime":"2025-08-11 07:39","pubTimestamp":1754869140,"startTime":"0","endTime":"0","summary":"有报道称，陈天桥对代季峰领衔的这家新 AI 创业公司寄予厚望，还承诺，盛大内部孵化的所有AI企业的一半利润将分给团队。如今，AI大牛代季峰再度“出山”，与创新企业家、慈善家、天桥脑科学研究院创始人陈天桥联手筹备一家新的AI创业公司，目标是打造下一个OpenAI，第二个DeepSeek，将围绕AGI展开基础性研究，首个项目就是MiroMind Open Deep Research。目前，代季峰的MiroMind团队已经对外开放MiroMind ODR项目Demo进行体验。","market":"fut","thumbnail":null,"type":0,"news_type":0,"thumbnails":[],"rights":null,"property":[],"language":"zh","translate_title":"","themeId":null,"theme_name":"","theme_type":"","isJumpTheme":false,"source_url":"https://finance.sina.com.cn/stock/t/2025-08-11/doc-infkqfti6088263.shtml","is_publish_highlight":false,"source_rank":0,"column":"","sentiment":"1","news_top_title":null,"news_tag":"","news_rank":0,"length":0,"strategy_id":0,"source":"sina","symbols":["BK4017","LU0368265764.SGD","LU0498741890.SGD","LU0055631609.USD","AGI","SFT","LU0498741114.HKD","BK4214","LU0496367417.USD"],"gpt_icon":0},{"id":"2557153503","title":"站在DeepSeek肩膀上，小紅書開源首款多模態模型：看懂表情包與數學題，一手實測","url":"https://stock-news.laohu8.com/highlight/detail?id=2557153503","media":"智东西","labels":[],"top":-1,"itemType":null,"share":"https://ttm.financial/m/news/2557153503?lang=zh_tw&edition=fundamental","pubTime":"2025-08-07 12:30","pubTimestamp":1754541040,"startTime":"0","endTime":"0","summary":"今年6月6日，小红书开源了其首款大语言模型，并在之后开源了用于OCR的专用模型，以及视觉、奖励模型等前沿方向的研究成果。VLM预训练在这一阶段，hi lab将视觉编码器与DeepSeek V3联合训练，使用大规模、多样化的多模态数据集，主要包括跨模态互译数据和跨模态融合数据。","market":"sh","thumbnail":null,"type":0,"news_type":0,"thumbnails":[],"rights":null,"property":[],"language":"zh","translate_title":"","themeId":null,"theme_name":"","theme_type":"","isJumpTheme":false,"source_url":"https://tech.ifeng.com/c/8lcepKwpv4H","is_publish_highlight":false,"source_rank":0,"column":"","sentiment":"0","news_top_title":null,"news_tag":"","news_rank":0,"length":0,"strategy_id":0,"source":"fenghuang_stock","symbols":["SFT","BK4214"],"gpt_icon":0},{"id":"2555003454","title":"DeepSeek V4借實習生獲獎論文「起飛」？梁文峯劍指上下文：處理速度提10倍、要「完美」準確率","url":"https://stock-news.laohu8.com/highlight/detail?id=2555003454","media":"AI前线","labels":[],"top":-1,"itemType":null,"share":"https://ttm.financial/m/news/2555003454?lang=zh_tw&edition=fundamental","pubTime":"2025-07-31 13:00","pubTimestamp":1753938027,"startTime":"0","endTime":"0","summary":"据了解，袁境阳当时写这篇论文时还只是 Deepseek 的实习生。上下文处理速度狂飙，准确率堪称“完美”在真实世界语言语料库上进行综合实验评估后，NSA 由于稀疏性过滤掉更多噪声，在基准测试中产生更好的准确率。在 64k 上下文长度下，前向速度提升高达 9.0 倍，反向速度提升高达 6.0 倍。早在今年 4 月，就有“DeepSeek R2 提前泄露”的传言在 AI 圈刷屏。","market":"us","thumbnail":null,"type":0,"news_type":0,"thumbnails":[],"rights":null,"property":[],"language":"zh","translate_title":"","themeId":null,"theme_name":"","theme_type":"","isJumpTheme":false,"source_url":"https://tech.ifeng.com/c/8lPiESohHbv","is_publish_highlight":false,"source_rank":0,"column":"","sentiment":"0","news_top_title":null,"news_tag":"","news_rank":0,"length":0,"strategy_id":0,"source":"fenghuang_stock","symbols":["BK4214","SFT"],"gpt_icon":0},{"id":"2554725200","title":"硬核「吵」了30分鐘：這場大模型圓桌，把AI行業的分歧説透了","url":"https://stock-news.laohu8.com/highlight/detail?id=2554725200","media":"机器之心","labels":[],"top":-1,"itemType":null,"share":"https://ttm.financial/m/news/2554725200?lang=zh_tw&edition=fundamental","pubTime":"2025-07-28 12:24","pubTimestamp":1753676662,"startTime":"0","endTime":"0","summary":"2025 年世界人工智能大会第二天，几位行业大佬“吵”起来了。他们“吵架”的热度，堪比盛夏的天气，直逼 40 度高温。随着这个观点的抛出，这场围绕大模型的圆桌论坛正式拉开帷幕。右边表明在“测试时间”增加时，模型的表现也会得到改善。从 2017 年至今，Transformer 统治 AI 领域已经八年之久。","market":"hk","thumbnail":null,"type":0,"news_type":0,"thumbnails":[],"rights":null,"property":[],"language":"zh","translate_title":"","themeId":null,"theme_name":"","theme_type":"","isJumpTheme":false,"source_url":"https://tech.ifeng.com/c/8lM6x5suQao","is_publish_highlight":false,"source_rank":0,"column":"","sentiment":"0","news_top_title":null,"news_tag":"","news_rank":0,"length":0,"strategy_id":0,"source":"fenghuang_stock","symbols":["LU2106854487.HKD","IE0004445015.USD","NVDY","IE00BJLML261.HKD","LU1261432733.SGD","SG9999014542.SGD","LU2461242641.AUD","LU2552382215.SGD","LU1951200564.SGD","LU2286300806.USD","3NVD.UK","LU0154236417.USD","LU0466842654.USD","LU0985320562.USD","NVD","NVDS.UK","LU0310800379.SGD","LU0069063385.USD","IE00BQXX3C00.GBP","NVIW.SI","NVD3.UK","NVDD","LU0444973449.USD","NVD2.UK","LU0072462426.USD","LU1069344957.HKD","LU1232071149.USD","NVDX","LU0077335932.USD","LU1814569148.SGD","LU0354030511.USD","SNVD.UK","NVDS","LU2430703251.USD","SFT","IE0004091025.USD","LU1923622614.USD","2NVD.UK","IE00BMPRXQ63.HKD","IE00BJTD4V19.USD","IE00B1BXHZ80.USD","LU0203201768.USD","LU2746668461.USD","NVDU","LU0868494617.USD","IE00BKPKM429.USD","LU0342679015.USD","LU0109392836.USD"],"gpt_icon":1},{"id":"2553261454","title":"復旦聯合南洋理工提出基於視覺Grounding的多輪強化學習框架MGPO","url":"https://stock-news.laohu8.com/highlight/detail?id=2553261454","media":"市场资讯","labels":[],"top":-1,"itemType":null,"share":"https://ttm.financial/m/news/2553261454?lang=zh_tw&edition=fundamental","pubTime":"2025-07-21 13:46","pubTimestamp":1753076760,"startTime":"0","endTime":"0","summary":"为解决这一问题，复旦大学、南洋理工大学的研究者提出一种基于视觉 Grounding 的多轮强化学习方法 MGPO，使 LMM 能在多轮交互中根据问题，自动预测关键区域坐标，裁剪子图像并整合历史上下文，最终实现高分辨率图像的精准推理。但传统视觉 Grounding 模型需依赖大量 Grounding 标注进行训练，而此类标注成本较高。","market":"us","thumbnail":null,"type":0,"news_type":0,"thumbnails":[],"rights":null,"property":[],"language":"zh","translate_title":"","themeId":null,"theme_name":"","theme_type":"","isJumpTheme":false,"source_url":"https://finance.sina.com.cn/stock/t/2025-07-21/doc-infhfezw4892690.shtml","is_publish_highlight":false,"source_rank":0,"column":"","sentiment":"0","news_top_title":null,"news_tag":"","news_rank":0,"length":0,"strategy_id":0,"source":"sina","symbols":["BK4202","BK4214","SFT","LU0006061336.USD","RL","BK4585","BK4588"],"gpt_icon":0},{"id":"2552443983","title":"AI打假AI，拿下SOTA丨廈大&騰訊優圖","url":"https://stock-news.laohu8.com/highlight/detail?id=2552443983","media":"市场资讯","labels":[],"top":-1,"itemType":null,"share":"https://ttm.financial/m/news/2552443983?lang=zh_tw&edition=fundamental","pubTime":"2025-07-20 15:00","pubTimestamp":1752994800,"startTime":"0","endTime":"0","summary":"具体方法是AIGI-HolmesAI生成图像检测方法，由厦门大学多媒体可信感知与高效计算教育部重点实验室和腾讯优图团队带来。团队针对上述问题，通过AIGI-Holmes给出解决方案。其中CLIP用于检测high-level semantic缺陷，而NPR则用于检测low-level artfacts，分别在Holmes-set上进行LoRA微调和全参微调。具体来说，团队在三个AIGI检测的数据集上评估了检测能力，包括AIGCDetect-Benchmark、AntiFakePrompt，并且额外采集了10种SOTA生成模型的图片构建了第三个benchmark，用于测试模型在未见过的生成方法上的泛化能力。","market":"sg","thumbnail":null,"type":0,"news_type":0,"thumbnails":[],"rights":null,"property":[],"language":"zh","translate_title":"","themeId":null,"theme_name":"","theme_type":"","isJumpTheme":false,"source_url":"https://finance.sina.com.cn/stock/t/2025-07-20/doc-infhcaau4129839.shtml","is_publish_highlight":false,"source_rank":0,"column":"","sentiment":"0","news_top_title":null,"news_tag":"","news_rank":0,"length":0,"strategy_id":0,"source":"sina","symbols":["SFT","BK4214"],"gpt_icon":0},{"id":"2552987443","title":"任務級獎勵提升AppAgent思考力，淘天提出Mobile-R1，3B模型超32B","url":"https://stock-news.laohu8.com/highlight/detail?id=2552987443","media":"市场资讯","labels":[],"top":-1,"itemType":null,"share":"https://ttm.financial/m/news/2552987443?lang=zh_tw&edition=fundamental","pubTime":"2025-07-20 14:45","pubTimestamp":1752993900,"startTime":"0","endTime":"0","summary":"结果Mobile-R1顺利地完成了这一任务。更为显著的是，Mobile-R1在所有基准中表现最佳，任务成功率达到49.40，比最优秀的baseline model高出将近20点。Mobile-R1在处理未见应用时表现出良好的泛化性，而其他模型在泛化能力上存在挑战。Mobile-R1的优异表现主要归功于Stage 3的训练，这一阶段有效增强了模型的鲁棒性和适应性。实验结果表明，Mobile-R1在所有指标上都超越了所有基准。","market":"fut","thumbnail":null,"type":0,"news_type":0,"thumbnails":[],"rights":null,"property":[],"language":"zh","translate_title":"","themeId":null,"theme_name":"","theme_type":"","isJumpTheme":false,"source_url":"https://finance.sina.com.cn/stock/t/2025-07-20/doc-infhcaau9104916.shtml","is_publish_highlight":false,"source_rank":0,"column":"","sentiment":"0","news_top_title":null,"news_tag":"","news_rank":0,"length":0,"strategy_id":0,"source":"sina","symbols":["SFT","BK4214"],"gpt_icon":0},{"id":"2552495787","title":"為什麼2025成了Agent落地元年？","url":"https://stock-news.laohu8.com/highlight/detail?id=2552495787","media":"虎嗅APP","labels":[],"top":-1,"itemType":null,"share":"https://ttm.financial/m/news/2552495787?lang=zh_tw&edition=fundamental","pubTime":"2025-07-18 18:20","pubTimestamp":1752834000,"startTime":"0","endTime":"0","summary":"而Agent就是AWS给出的答案。但如何低成本、高质量的重新做一遍，如何让Agent加速落地呢？不同于过去将最新的模型发布作为重磅亮点，这一次的峰会，Agentic AI 是唯一的关键词。那么为什么是今年？他们的出现，进一步带动了Agent在千行百业的落地。相对而言，Gartner的预测更加保守也更具普适性代表，到2028年，33%的企业软件将使用Agentic AI，15%的日常工作决策将由Agent自主完成。","market":"hk","thumbnail":null,"type":0,"news_type":0,"thumbnails":[],"rights":null,"property":[],"language":"zh","translate_title":"","themeId":null,"theme_name":"","theme_type":"","isJumpTheme":false,"source_url":"http://mp.weixin.qq.com/s?__biz=MTQzMjE1NjQwMQ==&mid=2656106191&idx=2&sn=6793fb3de10f7ae21d37c57116920747&chksm=67aeb1b0f63ead745a2bedf475378e40181e00517f74318c5f93602b7620eb8c561e2e9b77ee&scene=0#rd","is_publish_highlight":false,"source_rank":0,"column":"","sentiment":"0","news_top_title":null,"news_tag":"","news_rank":0,"length":0,"strategy_id":0,"source":"weixin_highlight","symbols":["USAW.SI","BK4554","LU0157215616.USD","LU2756315318.SGD","BK4585","IE0034235303.USD","IE00BWXC8680.SGD","IE000KEQY171.SGD","LU0861579265.USD","LU0267386448.USD","SGXZ31699556.SGD","IE00BQXX3C00.GBP","LU0276348264.USD","LU2764263203.CNY","LU0130102774.USD","LU0308772762.SGD","LU0345769128.USD","LU2505996509.AUD","LU0353189680.USD","CPT","LU1066051225.USD","LU2108987350.USD","LU0345768153.USD","LU0314104364.USD","LU0079474960.USD","LU1069344957.HKD","LU1116320901.HKD","LU1232071149.USD","LU0689472784.USD","IE000W1ABFV2.USD","LU2097829019.USD","LU1791710582.SGD","LU2505996681.GBP","LU1989764748.USD","LU0077335932.USD","LU0965509283.SGD","SG9999004303.SGD","LU0642271901.SGD","SG9999017495.SGD","IE00BKVL7J92.USD","SFT","SG9999018857.SGD","BK4596","LU0882574139.USD","LU0964807845.USD","LU2362540622.SGD","BK4215","IE00B19Z9505.USD","LU0203202063.USD"],"gpt_icon":0},{"id":"2551975105","title":"OpenAI聯合創始人揭秘AI進化新方向：讓模型學會和人類一樣反思","url":"https://stock-news.laohu8.com/highlight/detail?id=2551975105","media":"市场资讯","labels":[],"top":-1,"itemType":null,"share":"https://ttm.financial/m/news/2551975105?lang=zh_tw&edition=fundamental","pubTime":"2025-07-14 11:25","pubTimestamp":1752463500,"startTime":"0","endTime":"0","summary":"Karpathy 觉得，RL 缺少这种类似人类反思的机制，而这可能是 LLMs 未来进化的关键。Karpathy 用“second nature”来形容人类通过反思逐渐掌握技能的过程。Karpathy 认为，AI 应该也有类似机制，尤其是像 LLMs 这样有强大语言能力和上下文学习能力的模型。Karpathy 认为，RL 确实比监督微调更“苦涩”，而且还会带来更多性能提升。Karpathy 的设想是：如果能让模型自己总结经验教训，并在实践中不断优化，可能会开启 AI 智能的新篇章。","market":"fut","thumbnail":null,"type":0,"news_type":0,"thumbnails":[],"rights":null,"property":[],"language":"zh","translate_title":"","themeId":null,"theme_name":"","theme_type":"","isJumpTheme":false,"source_url":"https://finance.sina.com.cn/stock/t/2025-07-14/doc-inffmawy3285895.shtml","is_publish_highlight":false,"source_rank":0,"column":"","sentiment":"0","news_top_title":null,"news_tag":"","news_rank":0,"length":0,"strategy_id":0,"source":"sina","symbols":["LU0006061336.USD","BK4585","RL","BK4588","SFT","BK4214","BK4202"],"gpt_icon":0},{"id":"2551973092","title":"Karpathy戳破強化學習神話，首提AI覆盤式進化！暴力試錯將死","url":"https://stock-news.laohu8.com/highlight/detail?id=2551973092","media":"市场资讯","labels":[],"top":-1,"itemType":null,"share":"https://ttm.financial/m/news/2551973092?lang=zh_tw&edition=fundamental","pubTime":"2025-07-14 11:06","pubTimestamp":1752462360,"startTime":"0","endTime":"0","summary":"Karpathy最新发文提出另一种Scaling范式，像人类一样反思回顾，通过复盘学习取得突破，更多的S形进步曲线等待发现。然而， 在Karpathy看来，从长远角度来讲，强化学习或许并不是最优策略。一位网友很有见地称，强化学习实际上是暴力试错的一种方法，并非是明智的策略。放弃无效RL研究最近，关于强化学习的讨论，成为了AI圈的一大热点。除了Karpathy本人下场，上周前OpenAI研究员Kevin Lu发长文称，Transformer只是配角，放弃无效RL研究！","market":"sg","thumbnail":null,"type":0,"news_type":0,"thumbnails":[],"rights":null,"property":[],"language":"zh","translate_title":"","themeId":null,"theme_name":"","theme_type":"","isJumpTheme":false,"source_url":"https://finance.sina.com.cn/stock/t/2025-07-14/doc-inffmaxa5084189.shtml","is_publish_highlight":false,"source_rank":0,"column":"","sentiment":"0","news_top_title":null,"news_tag":"","news_rank":0,"length":0,"strategy_id":0,"source":"sina","symbols":["SFT","BK4214"],"gpt_icon":0},{"id":"2550614919","title":"豆蔻婦科大模型再突破：釘釘行業訓練平台+精標數據SFT ，準確率從 77.1%上升至 90.2%","url":"https://stock-news.laohu8.com/highlight/detail?id=2550614919","media":"市场资讯","labels":[],"top":-1,"itemType":null,"share":"https://ttm.financial/m/news/2550614919?lang=zh_tw&edition=fundamental","pubTime":"2025-07-10 15:39","pubTimestamp":1752133140,"startTime":"0","endTime":"0","summary":"豆蔻妇科大模型的模型调优经历了两个关键优化阶段：以下是豆蔻妇科大模型从第一个版本的准确率77.1%，通过进一步的SFT后，准确率达到90.2%我们团队的一些方法和心得，供大家参考，欢迎留言讨论。例如，针对 “患者出现阴道出血症状，诊断为宫颈病变” 等诊断结论，依据医学知识库中的关联规则，自动评估其逻辑合理性与临床可行性。在这一阶段的调优过程中，钉钉企业专属AI平","market":"other","thumbnail":null,"type":0,"news_type":0,"thumbnails":[],"rights":null,"property":[],"language":"zh","translate_title":"","themeId":null,"theme_name":"","theme_type":"","isJumpTheme":false,"source_url":"https://finance.sina.com.cn/stock/t/2025-07-10/doc-infeypam3762985.shtml","is_publish_highlight":false,"source_rank":0,"column":"","sentiment":"1","news_top_title":null,"news_tag":"","news_rank":0,"length":0,"strategy_id":0,"source":"sina","symbols":["LU0737861772.HKD","LU0499858602.USD","LU1051768304.USD","LU0593848301.USD","SFT","BK1521","LU0516423174.USD","LU0048580855.USD","LU0588546209.SGD","ALBmain","LU0293314216.USD","LU0345776255.USD","LU0348784397.USD","SG9999001689.USD","LU0823397285.USD","BK1584","BK1615","LU1568876335.HKD","IE0034224299.USD","LU0084288322.USD","LU0640798160.USD","LU0831103253.SGD","SG9999014674.SGD","LU0488056044.USD","89988","LU0348735423.USD","LU1515016050.SGD","HBBD.SI","LU0449515922.USD","LU2242644610.SGD","LU0315178854.USD","LU0029874905.USD","LU0029875118.USD","LU0348805143.USD","LU0140636845.USD","LU0163747925.USD","LU1201861165.SGD","LU0642271901.SGD","LU0329678337.USD","LU0819123356.HKD","BK1517","LU0449509016.USD","LU0979878070.USD","LU0611395673.USD","LU0577902371.SGD","LU1323998911.USD","LU0516423091.SGD","LU1808992512.USD"],"gpt_icon":1},{"id":"2550692415","title":"vivo發端側多模態模型，只有3B可理解GUI界面，20項評測表現亮眼","url":"https://stock-news.laohu8.com/highlight/detail?id=2550692415","media":"市场资讯","labels":[],"top":-1,"itemType":null,"share":"https://ttm.financial/m/news/2550692415?lang=zh_tw&edition=fundamental","pubTime":"2025-07-10 13:13","pubTimestamp":1752124380,"startTime":"0","endTime":"0","summary":"vivo AI Lab发布AI多模态新模型了，专门面向端侧设计，紧凑高效～能够直接理解GUI页面的那种：模型BlueLM-2.5-3B，融合文本和图文的理解和推理能力，支持长短思考模式自由切换，并引入思考预算控制机制。例如，在AIME25 任务中thinking模式较之non-thinking 模式提高达40分，在MathVision 任务中提高达19.2分。此外，小尺寸ViT也有助于进一步降低功耗。将文本任务的推理增强训练后置到多模态阶段，有效避免了文本推理能力遗忘，提升了训练效率。","market":"us","thumbnail":null,"type":0,"news_type":0,"thumbnails":[],"rights":null,"property":[],"language":"zh","translate_title":"","themeId":null,"theme_name":"","theme_type":"","isJumpTheme":false,"source_url":"https://finance.sina.com.cn/stock/t/2025-07-10/doc-infeyhuu0412326.shtml","is_publish_highlight":false,"source_rank":0,"column":"","sentiment":"0","news_top_title":null,"news_tag":"","news_rank":0,"length":0,"strategy_id":0,"source":"sina","symbols":["SFT","BK4214"],"gpt_icon":0},{"id":"2548865185","title":"智譜再融10億！獲上海國資押注，開源視覺大模型，能解説球賽，還會玩手機","url":"https://stock-news.laohu8.com/highlight/detail?id=2548865185","media":"市场资讯","labels":[],"top":-1,"itemType":null,"share":"https://ttm.financial/m/news/2548865185?lang=zh_tw&edition=fundamental","pubTime":"2025-07-02 15:40","pubTimestamp":1751442000,"startTime":"0","endTime":"0","summary":"开源之外，智谱还在今天举行的智谱开放平台产业生态大会上宣布，该公司获得浦东创投集团和张江集团联合战略投资，总额10亿元。目前，开源社区缺乏一种在广泛任务范围内持续超越传统同类参数规模非推理模型的多模态推理模型。在视觉编码器部分，智谱将原始的二维卷积替换为三维卷积，尤其适用于视频理解，有效提升了处理效率。","market":"other","thumbnail":null,"type":0,"news_type":0,"thumbnails":[],"rights":null,"property":[],"language":"zh","translate_title":"","themeId":null,"theme_name":"","theme_type":"","isJumpTheme":false,"source_url":"https://finance.sina.com.cn/stock/t/2025-07-02/doc-infeanhk2500975.shtml","is_publish_highlight":false,"source_rank":0,"column":"","sentiment":"1","news_top_title":null,"news_tag":"","news_rank":0,"length":0,"strategy_id":0,"source":"sina","symbols":["RL","LU0006061336.USD","BK4588","BK4202","BK4214","BK4585","SFT"],"gpt_icon":0},{"id":"2548597918","title":"10B級模型SOTA，超8倍參數「大」模型，智譜開源GLM-4.1V-Thinking","url":"https://stock-news.laohu8.com/highlight/detail?id=2548597918","media":"市场资讯","labels":[],"top":-1,"itemType":null,"share":"https://ttm.financial/m/news/2548597918?lang=zh_tw&edition=fundamental","pubTime":"2025-07-02 12:22","pubTimestamp":1751430120,"startTime":"0","endTime":"0","summary":"智谱正式发布 GLM-4.1V-Thinking 系列模型，并率先开源GLM-4.1V-9B-Thinking，标志着智谱 GLM 视觉大模型向高阶认知迈出了关键一步。在 18 项权威评测中，GLM-4.1V-9B-Thinking 的表现已可比肩甚至超越参数量高达 72B 的 Qwen2.5-VL-72B，充分展示出结构设计与训练策略的先进性与效率。模型原理1. 模型架构GLM-4.1V-Thinking 模型架构由三个核心模块组成：视觉编码器、多层感知机适配器以及语言解码器。2训练流程GLM-4.1V-Thinking 的训练过程分为三个阶段：预训练、监督微调 和 强化学习。","market":"sh","thumbnail":null,"type":0,"news_type":0,"thumbnails":[],"rights":null,"property":[],"language":"zh","translate_title":"","themeId":null,"theme_name":"","theme_type":"","isJumpTheme":false,"source_url":"https://finance.sina.com.cn/stock/t/2025-07-02/doc-infczzsq2682912.shtml","is_publish_highlight":false,"source_rank":0,"column":"","sentiment":"0","news_top_title":null,"news_tag":"","news_rank":0,"length":0,"strategy_id":0,"source":"sina","symbols":["BK4588","SFT","BK4585","RL","BK4214","BK4202","LU0006061336.USD"],"gpt_icon":0},{"id":"2548248960","title":"SuperCLUE推理榜驚現黑馬：原來中興是一家AI公司？","url":"https://stock-news.laohu8.com/highlight/detail?id=2548248960","media":"市场资讯","labels":[],"top":-1,"itemType":null,"share":"https://ttm.financial/m/news/2548248960?lang=zh_tw&edition=fundamental","pubTime":"2025-07-01 13:02","pubTimestamp":1751346120,"startTime":"0","endTime":"0","summary":"前段时间，中文大模型测评基准 SuperCLUE 发布了 2025 年 5 月报告。SuperCLUE 推理榜单深度聚焦模型的逻辑思维与问题解决能力，涵盖数学推理、科学推理、代码生成三大硬核维度。但是，星云大模型 NebulaCoder-V6 着实算得上一匹黑马，因为它来自一家老牌信息通信公司 —— 中兴通讯。屠嘉顺同意这种“6G 将是 AI 原生”的说法。此次星云大模型在 SuperCLUE 推理榜单夺冠，离不开技术团队设计的大模型高效训练优化方案。","market":"us","thumbnail":null,"type":0,"news_type":0,"thumbnails":[],"rights":null,"property":[],"language":"zh","translate_title":"","themeId":null,"theme_name":"","theme_type":"","isJumpTheme":false,"source_url":"https://finance.sina.com.cn/roll/2025-07-01/doc-infcxwtq3528857.shtml","is_publish_highlight":false,"source_rank":0,"column":"","sentiment":"1","news_top_title":null,"news_tag":"","news_rank":0,"length":0,"strategy_id":0,"source":"sina","symbols":["LU1571399168.USD","BK4018","LU1599440770.SGD","LU1815336091.USD","BK4532","LU1236620834.HKD","CRL","BK4214","LU2125154935.USD","SG9999001440.SGD","SFT","LU1236620750.USD","LU2360032135.SGD","LU1941712264.USD","LU1585245621.USD","LU2125154778.USD","LU0390134368.USD","LU0266013472.USD","LU0985320562.USD","LU1150488218.USD","BK4121","CL","LU2108987350.USD","BK4504","BK4533","LU1675838814.USD","BK4530","BK4585","LU0731783394.SGD","LU2463526074.USD","BK4588","LU1430594728.SGD","LU0757428866.USD","LU1941712348.USD","LU0323591593.USD","LU1150488135.SGD"],"gpt_icon":0}],"pageSize":20,"totalPage":2,"pageCount":1,"totalSize":31,"code":"91000000","status":"200"}]}}