智谱联合华为开源首个国产芯片训练的多模态SOTA模型

老虎资讯综合

Jan 14

智谱今天宣布联合华为开源新一代图像生成模型GLM-Image，模型基于昇腾Atlas 800T A2设备和昇思MindSpore AI框架完成从数据到训练的全流程，是首个在国产芯片上完成全程训练的SOTA多模态模型。

GLM-Image采用自主创新的「自回归+扩散解码器」混合架构，实现了图像生成与语言模型的联合，是其面向以Nano Banana Pro为代表的新一代「认知型生成」技术范式的一次重要探索。

核心亮点如下：

架构革新，面向「认知型生成」的技术探索：采用创新的「自回归 + 扩散编码器」混合架构，兼顾全局指令理解与局部细节刻画，克服了海报、PPT、科普图等知识密集型场景生成难题，向探索以Nano Banana Pro为代表的新一代“知识+推理”的认知型生成模型迈出了重要一步。
首个在国产芯片完成全程训练的SOTA模型：模型自回归结构基座基于昇腾Atlas 800T A2设备与昇思MindSpore AI框架，完成了从数据预处理到大规模训练的全流程构建，验证了在国产全栈算力底座上训练前沿模型的可行性。
文字渲染开源SOTA：在CVTG-2K（复杂视觉文本生成）和LongText-Bench（长文本渲染）榜单获得开源第一，尤其擅长汉字生成任务。
高性价比与速度优化：API调用模式下，生成一张图片仅需0.1元，速度优化版本即将更新。

Disclaimer: Investing carries risk. This is not financial advice. The above content should not be regarded as an offer, recommendation, or solicitation on acquiring or disposing of any financial products, any associated discussions, comments, or posts by author or other users should not be considered as such either. It is solely for general information purpose only, which does not consider your own investment objectives, financial situations or needs. TTM assumes no responsibility or warranty for the accuracy and completeness of the information, investors should do their own research and may seek professional advice before investing.

Most Discussed

1
2
3
4
5
6
7
8
9
10

{"basename":"","ssrTDKData":{"titleTemplate":"%s - Tiger Brokers","title":"Tiger Brokers | Global Stocks, Options & Futures Trading App","description":"Tiger Brokers, one-stop investment in US stocks, SGX stocks, HK stocks, A-shares & other global assets. One of the best stock trading platforms in Singapore.","keywords":"tiger brokers,tiger trade,tiger brokers singapore,broker online,stock trading in singapore,share trading singapore,brokerage firm singapore,trading app,stock broker singapore,stock trading platforms,trading account","social":{"ogDescription":"Tiger Brokers, one-stop investment in US stocks, SGX stocks, HK stocks, A-shares & other global assets. One of the best stock trading platforms in Singapore.","ogImage":"https://c1.itigergrowtha.com/portal5/static/media/og-logo.be62fbe1.png","ogUrl":"https://www.itiger.com/news/1120279175"},"companyName":"Tiger Brokers"},"pageData":{"isMobile":false,"isTiger":false,"isTTM":true,"region":"SGP","license":"TBSG","edition":"fundamental"},"isCrawlerRequest":true,"__swrFallback__":{"@#url:\"https://stock-news.skytigris.cn/v3/news\",params:#id:\"1120279175\",edition:\"fundamental\",auth_exemption:1,,,undefined,":{"share":"https://ttm.financial/m/news/1120279175?lang=en_US&edition=fundamental","thumbnail":"https://community-static.tradeup.com/news/619230baba896a31e5bc24123cbb201f","is_english":false,"pubTime":"2026-01-14 09:13","share_image_url":"https://static.laohu8.com/e9f99090a1c2ed51c021029395664489","id":"1120279175","market":"us","top_or_hot":-1,"title":"智谱联合华为开源首个国产芯片训练的多模态SOTA模型","media":"老虎资讯综合","content":"<html><head></head><body><p>智谱今天宣布联合华为开源新一代图像生成模型GLM-Image，模型基于昇腾Atlas 800T A2设备和昇思MindSpore AI框架完成从数据到训练的全流程，是首个在国产芯片上完成全程训练的SOTA多模态模型。</p><p>GLM-Image采用自主创新的「自回归+扩散解码器」混合架构，实现了图像生成与语言模型的联合，是其面向以Nano Banana Pro为代表的新一代「认知型生成」技术范式的一次重要探索。</p><p style=\"text-align: start;\">核心亮点如下：</p><ul style=\"\"><li><p><strong>架构革新，面向「认知型生成」的技术探索：</strong>采用创新的「自回归 + 扩散编码器」混合架构，兼顾全局指令理解与局部细节刻画，克服了海报、PPT、科普图等知识密集型场景生成难题，向探索以Nano Banana Pro为代表的新一代“知识+推理”的认知型生成模型迈出了重要一步。</p></li><li><p><strong>首个在国产芯片完成全程训练的SOTA模型：模型自回归结构基座</strong>基于昇腾Atlas 800T A2设备与昇思MindSpore AI框架，完成了从数据预处理到大规模训练的全流程构建，验证了在国产全栈算力底座上训练前沿模型的可行性。</p></li><li><p><strong>文字渲染开源SOTA：</strong>在CVTG-2K（复杂视觉文本生成）和LongText-Bench（长文本渲染）榜单获得开源第一，<strong>尤其擅长汉字生成任务</strong>。</p></li><li><p><strong>高性价比与速度优化：API调用模式下，生成一张图片仅需0.1元，速度优化版本即将更新。</strong></p></li></ul><p></p><p class=\"t-img-caption\"><img src=\"https://community-static.tradeup.com/news/50cdec16f19b92c6073352983c46edb7\" tg-width=\"1123\" tg-height=\"751\"/></p><p></p></body></html>","source":null,"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>智谱联合华为开源首个国产芯片训练的多模态SOTA模型</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智谱联合华为开源首个国产芯片训练的多模态SOTA模型\n</h2>\n\n<h4 class=\"meta\">\n\n\n<a class=\"head\" href=\"https://laohu8.com/wemedia/102\">\n\n\n<div class=\"h-thumb\" style=\"background-image:url(https://static.tigerbbs.com/8274c5b9d4c2852bfb1c4d6ce16c68ba);background-size:cover;\"></div>\n\n<div class=\"h-content\">\n<p class=\"h-name\">老虎资讯综合 </p>\n<p class=\"h-time\">2026-01-14 09:13</p>\n</div>\n\n</a>\n\n\n</h4>\n\n</header>\n<article>\n<html><head></head><body><p>智谱今天宣布联合华为开源新一代图像生成模型GLM-Image，模型基于昇腾Atlas 800T A2设备和昇思MindSpore AI框架完成从数据到训练的全流程，是首个在国产芯片上完成全程训练的SOTA多模态模型。</p><p>GLM-Image采用自主创新的「自回归+扩散解码器」混合架构，实现了图像生成与语言模型的联合，是其面向以Nano Banana Pro为代表的新一代「认知型生成」技术范式的一次重要探索。</p><p style=\"text-align: start;\">核心亮点如下：</p><ul style=\"\"><li><p><strong>架构革新，面向「认知型生成」的技术探索：</strong>采用创新的「自回归 + 扩散编码器」混合架构，兼顾全局指令理解与局部细节刻画，克服了海报、PPT、科普图等知识密集型场景生成难题，向探索以Nano Banana Pro为代表的新一代“知识+推理”的认知型生成模型迈出了重要一步。</p></li><li><p><strong>首个在国产芯片完成全程训练的SOTA模型：模型自回归结构基座</strong>基于昇腾Atlas 800T A2设备与昇思MindSpore AI框架，完成了从数据预处理到大规模训练的全流程构建，验证了在国产全栈算力底座上训练前沿模型的可行性。</p></li><li><p><strong>文字渲染开源SOTA：</strong>在CVTG-2K（复杂视觉文本生成）和LongText-Bench（长文本渲染）榜单获得开源第一，<strong>尤其擅长汉字生成任务</strong>。</p></li><li><p><strong>高性价比与速度优化：API调用模式下，生成一张图片仅需0.1元，速度优化版本即将更新。</strong></p></li></ul><p></p><p class=\"t-img-caption\"><img src=\"https://community-static.tradeup.com/news/50cdec16f19b92c6073352983c46edb7\" tg-width=\"1123\" tg-height=\"751\"/></p><p></p></body></html>\n\n</article>\n</div>\n</body>\n</html>\n","isBrief":false,"type":0,"news_type":1,"symbol":"02513","symbol_name":"智谱","start_time":0,"source_url":"","article_id":"1120279175","we_media_id":"102","thumbnails":["https://community-static.tradeup.com/news/619230baba896a31e5bc24123cbb201f"],"rights":null,"url":"https://stock-news.laohu8.com/highlight/detail?id=1120279175","pubTimestamp":1768353203,"columns":[],"sourceInfo":null,"weMediaInfo":{"media_name":"老虎资讯综合","introduction":"为用户提供金融资讯、行情、数据，旨在帮助投资者理解世界，做投资决策。","home_visible":1,"id":"102","head_image":"https://static.tigerbbs.com/8274c5b9d4c2852bfb1c4d6ce16c68ba"},"summary":"智谱今天宣布联合华为开源新一代图像生成模型GLM-Image，模型基于昇腾Atlas 800T A2设备和昇思MindSpore AI框架完成从数据到训练的全流程，是首个在国产芯片上完成全程训练的SOTA多模态模型。GLM-Image采用自主创新的「自回归+扩散解码器」混合架构，实现了图像生成与语言模型的联合，是我们面向以Nano Banana Pro为代表的新一代「认知型生成」技术范式的一次重要探索。文字渲染开源SOTA：在CVTG-2K和LongText-Bench榜单获得开源第一，尤其擅长汉字生成任务。","collect":0,"end_time":0,"defaultTopTitle":"","property":[],"viewcount":null,"language":"zh","relate_stocks":{"02513":"智谱"},"translate_title":"Zhipu and Huawei open source the first multi-modal SOTA model trained by domestic chips","themeId":"f7942cd83e7032c83d590e3a3cbe5ca5","isJumpTheme":false,"ttsUrl":"https://static.tigerbbs.com/e46a6679695190f61f965afb541b6040","symbols_score_info":{"02513":2},"content_text":"智谱今天宣布联合华为开源新一代图像生成模型GLM-Image，模型基于昇腾Atlas 800T A2设备和昇思MindSpore AI框架完成从数据到训练的全流程，是首个在国产芯片上完成全程训练的SOTA多模态模型。GLM-Image采用自主创新的「自回归+扩散解码器」混合架构，实现了图像生成与语言模型的联合，是其面向以Nano Banana Pro为代表的新一代「认知型生成」技术范式的一次重要探索。核心亮点如下：架构革新，面向「认知型生成」的技术探索：采用创新的「自回归 + 扩散编码器」混合架构，兼顾全局指令理解与局部细节刻画，克服了海报、PPT、科普图等知识密集型场景生成难题，向探索以Nano Banana Pro为代表的新一代“知识+推理”的认知型生成模型迈出了重要一步。首个在国产芯片完成全程训练的SOTA模型：模型自回归结构基座基于昇腾Atlas 800T A2设备与昇思MindSpore AI框架，完成了从数据预处理到大规模训练的全流程构建，验证了在国产全栈算力底座上训练前沿模型的可行性。文字渲染开源SOTA：在CVTG-2K（复杂视觉文本生成）和LongText-Bench（长文本渲染）榜单获得开源第一，尤其擅长汉字生成任务。高性价比与速度优化：API调用模式下，生成一张图片仅需0.1元，速度优化版本即将更新。","kind":"news","is_publish_news":true,"is_publish_highlight":true,"is_publish_live":false,"is_publish_wemedia":null,"editions":null,"column":"","sentiment":"1","news_tag":"corporation","news_rank":0,"symbols":[],"gpt_button":1,"need_auth":false,"code":"91000000","status":"200"}}}