谷歌首款混合推理Gemini 2.5登场，成本暴降600%！思考模式一开，直追o4-mini

新智元2025-04-18

编辑：桃子好困【新智元导读】谷歌发布首款混合推理模型Gemini 2.5 Flash，引入了革命性“思考预算”，可灵活控制推理深度，性能一举击败Claude 3.7，比肩o4-mini。而且，关闭思考模式成本直降600%。刚刚，谷歌重磅发布首个混合推理模型——Gemini 2.5 Flash。与Claude类似，新模型的“思考预算”可以自定义，即可开启/关闭Gemini 2.5的思考模式。...

网页链接

免责声明：本文观点仅代表作者个人观点，不构成本平台的投资建议，本平台不对文章信息准确性、完整性和及时性做出任何保证，亦不对因使用或信赖文章信息引发的任何损失承担责任。

精彩评论

我们需要你的真知灼见来填补这片空白

发表看法

{"i18n":{"language":"zh_CN"},"isChannel":false,"data":{"share":"https://www.laohu8.com/m/news/2528034687?lang=zh_CN&edition=full","thumbnail":"","is_english":false,"pubTime":"2025-04-18 09:50","share_image_url":"https://static.laohu8.com/e9f99090a1c2ed51c021029395664489","id":"2528034687","market":"us","top_or_hot":-1,"title":"谷歌首款混合推理Gemini 2.5登场，成本暴降600%！思考模式一开，直追o4-mini","media":"新智元","content":"<div>\n<p>编辑：桃子 好困【新智元导读】谷歌发布首款混合推理模型Gemini 2.5 Flash，引入了革命性“思考预算”，可灵活控制推理深度，性能一举击败Claude 3.7，比肩o4-mini。而且，关闭思考模式成本直降600%。刚刚，谷歌重磅发布首个混合推理模型——Gemini 2.5 Flash。与Claude类似，新模型的“思考预算”可以自定义，即可开启/关闭Gemini 2.5的思考模式。...</p>\n\n<a href=\"https://tech.ifeng.com/c/8ieHUBuoOSN\">网页链接</a>\n\n</div>\n","source":"fenghuang_stock","html":"<!DOCTYPE html>\n<html>\n<head>\n<meta http-equiv=\"Content-Type\" content=\"text/html; charset=utf-8\" />\n<meta name=\"viewport\" content=\"width=device-width,initial-scale=1.0,minimum-scale=1.0,maximum-scale=1.0,user-scalable=no\"/>\n<meta name=\"format-detection\" content=\"telephone=no,email=no,address=no\" />\n<title>谷歌首款混合推理Gemini 2.5登场，成本暴降600%！思考模式一开，直追o4-mini</title>\n<style type=\"text/css\">\na,abbr,acronym,address,applet,article,aside,audio,b,big,blockquote,body,canvas,caption,center,cite,code,dd,del,details,dfn,div,dl,dt,\nem,embed,fieldset,figcaption,figure,footer,form,h1,h2,h3,h4,h5,h6,header,hgroup,html,i,iframe,img,ins,kbd,label,legend,li,mark,menu,nav,\nobject,ol,output,p,pre,q,ruby,s,samp,section,small,span,strike,strong,sub,summary,sup,table,tbody,td,tfoot,th,thead,time,tr,tt,u,ul,var,video{ font:inherit;margin:0;padding:0;vertical-align:baseline;border:0 }\nbody{ font-size:16px; line-height:1.5; color:#999; background:transparent; }\n.wrapper{ overflow:hidden;word-break:break-all;padding:10px; }\nh1,h2{ font-weight:normal; line-height:1.35; margin-bottom:.6em; }\nh3,h4,h5,h6{ line-height:1.35; margin-bottom:1em; }\nh1{ font-size:24px; }\nh2{ font-size:20px; }\nh3{ font-size:18px; }\nh4{ font-size:16px; }\nh5{ font-size:14px; }\nh6{ font-size:12px; }\np,ul,ol,blockquote,dl,table{ margin:1.2em 0; }\nul,ol{ margin-left:2em; }\nul{ list-style:disc; }\nol{ list-style:decimal; }\nli,li p{ margin:10px 0;}\nimg{ max-width:100%;display:block;margin:0 auto 1em; }\nblockquote{ color:#B5B2B1; border-left:3px solid #aaa; padding:1em; }\nstrong,b{font-weight:bold;}\nem,i{font-style:italic;}\ntable{ width:100%;border-collapse:collapse;border-spacing:1px;margin:1em 0;font-size:.9em; }\nth,td{ padding:5px;text-align:left;border:1px solid #aaa; }\nth{ font-weight:bold;background:#5d5d5d; }\n.symbol-link{font-weight:bold;}\n/* header{ border-bottom:1px solid #494756; } */\n.title{ margin:0 0 8px;line-height:1.3;color:#ddd; }\n.meta {color:#5e5c6d;font-size:13px;margin:0 0 .5em; }\na{text-decoration:none; color:#2a4b87;}\n.meta .head { display: inline-block; overflow: hidden}\n.head .h-thumb { width: 30px; height: 30px; margin: 0; padding: 0; border-radius: 50%; float: left;}\n.head .h-content { margin: 0; padding: 0 0 0 9px; float: left;}\n.head .h-name {font-size: 13px; color: #eee; margin: 0;}\n.head .h-time {font-size: 11px; color: #7E829C; margin: 0;line-height: 11px;}\n.small {font-size: 12.5px; display: inline-block; transform: scale(0.9); -webkit-transform: scale(0.9); transform-origin: left; -webkit-transform-origin: left;}\n.smaller {font-size: 12.5px; display: inline-block; transform: scale(0.8); -webkit-transform: scale(0.8); transform-origin: left; -webkit-transform-origin: left;}\n.bt-text {font-size: 12px;margin: 1.5em 0 0 0}\n.bt-text p {margin: 0}\n</style>\n</head>\n<body>\n<div class=\"wrapper\">\n<header>\n<h2 class=\"title\">\n谷歌首款混合推理Gemini 2.5登场，成本暴降600%！思考模式一开，直追o4-mini\n</h2>\n\n<h4 class=\"meta\">\n\n\n2025-04-18 09:50 北京时间&nbsp;&nbsp;&nbsp;<a href=https://tech.ifeng.com/c/8ieHUBuoOSN><strong>新智元</strong></a>\n\n\n</h4>\n\n</header>\n<article>\n<div>\n<p>编辑：桃子 好困【新智元导读】谷歌发布首款混合推理模型Gemini 2.5 Flash，引入了革命性“思考预算”，可灵活控制推理深度，性能一举击败Claude 3.7，比肩o4-mini。而且，关闭思考模式成本直降600%。刚刚，谷歌重磅发布首个混合推理模型——Gemini 2.5 Flash。与Claude类似，新模型的“思考预算”可以自定义，即可开启/关闭Gemini 2.5的思考模式。...</p>\n\n<a href=\"https://tech.ifeng.com/c/8ieHUBuoOSN\">网页链接</a>\n\n</div>\n\n\n</article>\n</div>\n</body>\n</html>\n","isBrief":false,"type":0,"news_type":1,"symbol":"SG9999001077.SGD","symbol_name":"United International Growth Fund SGD","start_time":0,"source_url":"https://tech.ifeng.com/c/8ieHUBuoOSN","article_id":"2528034687","we_media_id":null,"thumbnails":[],"rights":null,"url":"https://stock-news.laohu8.com/highlight/detail?id=2528034687","pubTimestamp":1744941026,"columns":[],"sourceInfo":{"source_id":"fenghuang_stock","name":"凤凰网"},"weMediaInfo":null,"summary":"编辑：桃子 好困谷歌发布首款混合推理模型Gemini 2.5 Flash，引入了革命性“思考预算”，可灵活控制推理深度，性能一举击败Claude 3.7，比肩o4-mini。刚刚，谷歌重磅发布首个混合推理模型——Gemini 2.5 Flash。与Claude类似，新模型的“思考预算”可以自定义，即可开启/关闭Gemini 2.5的思考模式。值得一提的是，关闭思考的成本直接暴降600%，而且性能还不输Gemini 2.0 Flash。具体来说，Gemini 2.5 Flash关闭思考输出价格0.6美元/百万token，开启思考输出价格3.5美元/百万token。","collect":0,"end_time":0,"defaultTopTitle":"ifeng.com","property":[],"viewcount":null,"language":"zh","relate_stocks":{"SG9999001077.SGD":"United International Growth Fund SGD","LU2602419157.SGD":"HSBC ISLAMIC GLOBAL EQUITY INDEX \"AC\" (SGD) ACC","GOOG":"谷歌","LU1868837300.USD":"CT (LUX) I AMERICAN FUND \"9\" (USD) ACC","LU0345768153.USD":"NINETY ONE GSF GLOBAL STRATEGIC MANAGED \"A\" (USD) ACC","LU1496350502.SGD":"FRANKLIN DIVERSIFIED DYNAMIC \"A\" (SGDHDG) ACC","LU0353189680.USD":"富国美国全盘成长基金Cl A Acc","LU2362541513.USD":"WELLINGTON NEXT GENERATION GLOBAL EQUITY \"A\" (USD) ACC","LU0203201768.USD":"AB SICAV I - ALL MARKET INCOME PORTFOLIO \"AX\" (USD) INC","GOOGL":"谷歌A","LU1814569148.SGD":"WELLINGTON GLOBAL QUALITY GROWTH \"D\" (SGDHDG) ACC","LU2023250504.SGD":"Allianz Thematica Cl AMg DIS H2-SGD","LU1720051108.HKD":"ALLIANZ GLOBAL ARTIFICIAL INTELLIGENCE \"AT\" (HKD) ACC","LU0889566641.SGD":"FTSF - Templeton Shariah Global Equity A Acc SGD","LU1914381329.SGD":"Allianz Best Styles Global Equity Cl ET Acc H2-SGD","LU2065171311.SGD":"M&G (LUX) GLOBAL MAXIMA \"A\" (SGD) ACC","LU1244550221.USD":"FRANKLIN GLOBAL MULTI-ASSET  INCOME \"A\" (USDHEDGED) INC (M)","LU1674673428.USD":"HSBC GIF GLOBAL LOWER CARBON EQUITY \"AC\" (USD) ACC","USJW.SI":"ALPHAB 3xLongSG261006","LU1116320737.USD":"BGF SYSTEMATIC GLOBAL ENHANCED EQUITY YIELD \"A6\" (USD) INC","LU1069344957.HKD":"AB SICAV I - AMERICAN GROWTH PORTFOLIO \"AD\" (HKD) INC","LU1127390331.HKD":"AB SICAV I - ALL MARKET INCOME PORTFOLIO \"A\" (HKD) ACC","LU0312595415.SGD":"Schroder ISF Global Climate Change Equity A Acc SGD","LU0154236417.USD":"BGF US FLEXIBLE EQUITY \"A2\" ACC","LU1551013342.USD":"Allianz Income and Growth Cl AMg2 DIS USD","IE00BJTD4N35.SGD":"Neuberger Berman US Long Short Equity A1  Acc SGD-H","LU2471134879.HKD":"INVESCO GLOBAL EQUITY INCOME ADVANTAGE \"A\" (HKD) INC","SG9999014880.SGD":"大华全球优质成长基金Acc SGD","LU2168564222.USD":"AZ ALLOCATION - TREND \"AAZ\" (USDHDG) ACC","LU0289941410.SGD":"AB FCP I Dynamic Diversified AX SGD","LU2108987350.USD":"UBS (LUX) EQUITY SICAV GLOBAL OPPORTUNITY SUSTAINABLE (USD) \"P\" (USD) ACC","IE00B1XK9C88.USD":"PINEBRIDGE US LARGE CAP RESEARCH ENHANCED \"A\" (USD) ACC","LU0964807845.USD":"ALLIANZ INCOME & GROWTH \"A\" (USD) INC","BK4543":"AI","LU0107464264.USD":"abrdn SICAV I - GLOBAL INNOVATION EQUITY \"A\" (USD) ACC","LU1629891620.HKD":"ALLIANZ INCOME AND GROWTH \"AMG2\" (H2-HKD) INC","LU0980610538.SGD":"Natixis Harris Associates US Equity RA SGD-H","LU0215105999.USD":"SCHRODER ISF GLOBAL EQUITY \"A\" ACC","BK4077":"互动媒体与服务","LU1267930227.SGD":"TEMPLETON GLOBAL BALANCED \"AS\" (SGD) ACC A","LU0096364046.USD":"CT (LUX) I AMERICAN \"DU\" (USD) ACC","LU2720916845.USD":"BGF GLOBAL UNCONSTRAINED EQUITY \"A2\" (USD) ACC","LU2471134796.USD":"INVESCO GLOBAL EQUITY INCOME ADVANTAGE \"A\" (USD) INC","IE00BSNM7G36.USD":"NEUBERGER BERMAN SYSTEMATIC GLOBAL SUSTAINABLE VALUE \"A\" (USD) ACC","LU2361045086.USD":"WELLINGTON US QUALITY GROWTH \"A\" (USD) ACC","LU0683600562.USD":"AB SELECT US EQUITY \"A\" (USD) ACC","LU0433182093.SGD":"First Eagle Amundi International AS-C SGD","LU0786609619.USD":"高盛全球千禧一代股票组合Acc","BK4533":"AQR资本管理(全球第二大对冲基金)","LU0965509010.AUD":"AB LOW VOLATILITY EQUITY PORTFOLIO \"AD\" (AUDHDG) INC"},"translate_title":"Google's first hybrid reasoning Gemini 2.5 debuts, and the cost drops by 600%! As soon as the thinking mode is turned on, catch up with o4-mini","themeId":null,"isJumpTheme":false,"ttsUrl":null,"symbols_score_info":{"USJW.SI":0.6,"GOOGL":1,"GOOG":1},"content_text":"编辑：桃子 好困【新智元导读】谷歌发布首款混合推理模型Gemini 2.5 Flash，引入了革命性“思考预算”，可灵活控制推理深度，性能一举击败Claude 3.7，比肩o4-mini。而且，关闭思考模式成本直降600%。刚刚，谷歌重磅发布首个混合推理模型——Gemini 2.5 Flash。与Claude类似，新模型的“思考预算”可以自定义，即可开启/关闭Gemini 2.5的思考模式。值得一提的是，关闭思考的成本直接暴降600%，而且性能还不输Gemini 2.0 Flash。具体来说，Gemini 2.5 Flash关闭思考输出价格0.6美元/百万token，开启思考输出价格3.5美元/百万token。当然了，思考越久，模型性能也会随之提升。在GPQA知识问答中，新模型24k思考预算，性能提升了6%；对于代码任务（LiveCodeBench），16k思考预算性能最佳。\n\n\n\n\n\n\n在多项基准测试中，Gemini 2.5 Flash再次刷新SOTA。在大模型排行榜中，Flash预览版以1392 ELO高分位居第二，与GPT-4.5-preview、Grok 3并驾齐驱。在数学（AIME 2025/2024）、多模态推理（MMMU）、知识问答（GPQA）等基准上，Gemini 2.5 Flash完全碾压Claude 3.7 Sonnet，足以与最新o4-mini相媲美。就模型每百万token输入/输出价格来看，Gemini 2.5 Flash更具性价比。在人类最后一次考试中，Gemini 2.5 Flash拿下12.1%高分，仅次于o4-mini目前，Flash预览版可以在Gemini中使用，API同时向开发者开放。首款混合推理Gemini登场击败Claude 3.7混合推理模型，就是专为需要在性能、成本、延迟之间找到完美平衡的开发者而设计。Gemini 2.5 Flash不仅继承了2.0 Flash的高速响应特点，还引入了革命性的“思考模式”——可根据任务需求灵活调整推理深度。Gemini 2.5系是“思考模型”，能够在回答前先行推理。模型不会立即输出结果，而是先执行“思考”流程，更好地理解提示词，拆解复杂任务并规划回答。如下图所示，相较于2.0 Flash，Gemini 2.5 Flash在复杂任务，如数学推理、科研分析中表现更优异。在LMArena其他评估中，比如Hard Prompts、编码、长查询，Gemini 2.5 Flash全部拿下第一。另外从下图中可看出，在同类模型中，2.5 Flash以超高性价比领跑，兼具最优性能和极低成本的优势。网友实测在网友的实测中，2.5 Flash物理模拟能力足够惊艳，小球会随着多边形变化精准运动。而且，2.5 Flash还轻轻松松通过了4o-mini/o3无法通过的Galton Board（高尔顿板）测试。它还能根据精灵图，创建出自定义游戏城房间。另一位网友用了最大24k预算，让2.5 Flash设计出了一个《创：战纪》风格的游戏。提示：Create Design a visually striking Tron-style game in a single HTML file, where AI-controlled light cycles compete in fast-paced, strategic battles against each other如今Claude 3.7已经完全没有优势了，在设计登录界面时，Gemini 2.5 Flash用时最短速度最快。","kind":"news","is_publish_news":true,"is_publish_highlight":false,"is_publish_live":false,"is_publish_wemedia":null,"editions":null,"column":"","sentiment":"1","news_tag":"","news_rank":0,"symbols":[],"gpt_button":0,"need_auth":false,"code":"91000000","status":"200"},"commentList":[],"isCommentEnd":true,"newsSizeData":{"likeSize":0,"commentSize":0,"repostSize":0,"favoriteSize":0,"likeStatus":false,"favoriteStatus":false},"APP":{"userAgent":"Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; ClaudeBot/1.0; +claudebot@anthropic.com)","isDev":false,"isTTM":false,"tenantId":"TBCN","deviceId":"web-server-community-laohu8-v3","version":"4.36.0","shortVersion":"4.36.0","platform":"web","vendor":"web","appName":"laohu8","isIOS":false,"isAndroid":false,"isTiger":false,"isTHS":false,"isWeiXin":false,"isWeiXinMini":false,"isWeiBo":false,"isQQ":false,"isBaiduSwan":false,"isBaiduBox":false,"isDingTalk":false,"isToutiao":false,"isOnePlus":false,"isHuaWei":false,"isXiaomi":false,"isXiaomiWebView":false,"isOppo":false,"isVivo":false,"isSamsung":false,"isMobile":false},"href":"/m/news/2528034687","isCrawlerRequest":true}