{"id":30365,"date":"2025-03-10T10:59:38","date_gmt":"2025-03-10T02:59:38","guid":{"rendered":"https:\/\/www.1ai.net\/?p=30365"},"modified":"2025-03-10T10:59:38","modified_gmt":"2025-03-10T02:59:38","slug":"%e8%b6%85%e8%b6%8a-mistral-%e5%92%8c-qwen%ef%bc%9a%e8%b0%b7%e6%ad%8c-gemini-embedding-%e7%99%bb%e9%a1%b6-mteb%ef%bc%8c%e9%97%ae%e9%bc%8e%e6%9c%80%e5%bc%ba%e6%96%87%e6%9c%ac%e5%b5%8c%e5%85%a5-ai","status":"publish","type":"post","link":"https:\/\/www.1ai.net\/en\/30365.html","title":{"rendered":"Beyond Mistral and Qwen: Google Gemini Embedding Tops MTEB, Takes Top Spot as Strongest Text Embedding AI Model"},"content":{"rendered":"<p>3 \u6708 10 \u65e5\u6d88\u606f\uff0c<a href=\"https:\/\/www.1ai.net\/en\/tag\/%e8%b0%b7%e6%ad%8c\" title=\"\u3010\u67e5\u770b\u542b\u6709[\u8c37\u6b4c]\u6807\u7b7e\u7684\u6587\u7ae0\u3011\" target=\"_blank\" rel=\"noopener\">\u8c37\u6b4c<\/a>\u516c\u53f8\u4e8e 3 \u6708 7 \u65e5\u53d1\u5e03\u535a\u6587\uff0c\u5ba3\u5e03\u63a8\u51fa\u00a0<strong><a href=\"https:\/\/www.1ai.net\/en\/tag\/gemini-embedding\" title=\"\u3010\u67e5\u770b\u542b\u6709[Gemini Embedding]\u6807\u7b7e\u7684\u6587\u7ae0\u3011\" target=\"_blank\" rel=\"noopener\">Gemini Embedding<\/a><\/strong>\uff0c\u662f\u4e00\u6b3e\u57fa\u4e8e AI \u7684\u6587\u672c\u5904\u7406\u6a21\u578b\uff0c\u73b0\u5df2\u96c6\u6210\u81f3 Gemini API\u3002<\/p>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-30366\" title=\"26ca594dj00ssw0a5004nd000sg00hsp\" src=\"https:\/\/www.1ai.net\/wp-content\/uploads\/2025\/03\/26ca594dj00ssw0a5004nd000sg00hsp.jpg\" alt=\"26ca594dj00ssw0a5004nd000sg00hsp\" width=\"1024\" height=\"640\" \/><\/p>\n<p>\u8be5\u6a21\u578b\u5728 Massive Text Embedding Benchmark\uff08MTEB\uff09\u4e2d\u62d4\u5f97\u5934\u7b79\uff0c\u8d85\u8d8a\u4e86 <a href=\"https:\/\/www.1ai.net\/en\/tag\/mistral\" title=\"\u3010\u67e5\u770b\u542b\u6709[Mistral]\u6807\u7b7e\u7684\u6587\u7ae0\u3011\" target=\"_blank\" rel=\"noopener\">Mistral<\/a>\u3001Cohere \u548c <a href=\"https:\/\/www.1ai.net\/en\/tag\/qwen\" title=\"\u3010\u67e5\u770b\u542b\u6709[Qwen]\u6807\u7b7e\u7684\u6587\u7ae0\u3011\" target=\"_blank\" rel=\"noopener\">Qwen<\/a> \u7b49\u7ade\u4e89\u5bf9\u624b\uff0c\u6210\u4e3a\u5f53\u524d\u6027\u80fd\u6700\u5f3a\u7684\u6587\u672c\u5d4c\u5165\u6a21\u578b\u3002<\/p>\n<p>Gemini Embedding \u5c06\u6587\u672c\u8f6c\u6362\u4e3a\u6570\u503c\u8868\u793a\uff08\u5411\u91cf\uff09\uff0c\u4ece\u800c\u652f\u6301\u8bed\u4e49\u641c\u7d22\u3001\u63a8\u8350\u7cfb\u7edf\u548c\u6587\u6863\u68c0\u7d22\u7b49\u529f\u80fd\u3002\u5b83\u5728 MTEB \u57fa\u51c6\u6d4b\u8bd5\u4e2d\u8868\u73b0\u51fa\u8272\uff0c\u5e73\u5747\u4efb\u52a1\u5f97\u5206\u4e3a 68.32\uff0c\u663e\u8457\u9ad8\u4e8e Linq-Embed-Mistral \u548c gte-Qwen2-7B-instruct \u7b49\u6a21\u578b\uff0c\u8fbe\u5230 State-of-the-art\u3002<\/p>\n<p>State-of-the-art\uff08SOTA\uff09AI \u6a21\u578b\u662f\u6307\u5728\u7279\u5b9a\u4efb\u52a1\u6216\u9886\u57df\u4e2d\uff0c\u5f53\u524d\u8868\u73b0\u6700\u4f18\u7684\u6a21\u578b\u6216\u65b9\u6cd5\u3002\u8fd9\u4e9b\u6a21\u578b\u901a\u5e38\u901a\u8fc7\u5728\u5404\u79cd\u57fa\u51c6\u6d4b\u8bd5\u4e2d\u53d6\u5f97\u6700\u9ad8\u5206\u6765\u8bc1\u660e\u5176\u4f18\u8d8a\u6027\uff0c\u5e76\u4e14\u5f80\u5f80\u5728\u51c6\u786e\u6027\u3001\u6548\u7387\u6216\u80fd\u529b\u4e0a\u8d85\u8d8a\u4e86\u4e4b\u524d\u7684\u6a21\u578b\uff0c\u751a\u81f3\u5728\u67d0\u4e9b\u4efb\u52a1\u4e2d\u8fbe\u5230\u4e86\u4eba\u7c7b\u6c34\u5e73\u7684\u8868\u73b0\u3002<\/p>\n<p>\u8be5\u6a21\u578b\u5728\u914d\u5bf9\u5206\u7c7b\u4e0a\u5f97\u5206\u4e3a 85.13\uff1b\u5728\u68c0\u7d22\u4e0a\u5f97\u5206 67.71\uff0c\u5728\u91cd\u6392\u5e8f\u5f97\u5206\u4e3a 65.58\uff0c\u8868\u660e Gemini Embedding \u5728 AI \u641c\u7d22\u5f15\u64ce\u3001\u6587\u6863\u5206\u6790\u548c\u804a\u5929\u673a\u5668\u4eba\u4f18\u5316\u7b49\u5b9e\u9645\u5e94\u7528\u4e2d\u5177\u6709\u663e\u8457\u4f18\u52bf\u3002<\/p>\n<p>MTEB \u7531 Hugging Face \u521b\u5efa\uff0c\u901a\u8fc7 50 \u591a\u4e2a\u6570\u636e\u96c6\u8bc4\u4f30 AI \u6a21\u578b\u5728\u6587\u672c\u6570\u636e\u6392\u540d\u3001\u5206\u7c7b\u548c\u68c0\u7d22\u65b9\u9762\u7684\u80fd\u529b\u3002\u4f5c\u4e3a\u884c\u4e1a\u6807\u51c6\uff0cMTEB \u6392\u884c\u699c\u4e3a\u4f01\u4e1a\u5728\u9009\u62e9 AI \u6a21\u578b\u65f6\u63d0\u4f9b\u4e86\u91cd\u8981\u53c2\u8003\u3002Gemini Embedding \u7684\u4f18\u5f02\u8868\u73b0\u4e0d\u4ec5\u5de9\u56fa\u4e86\u8c37\u6b4c\u5728 AI \u9886\u57df\u7684\u9886\u5bfc\u5730\u4f4d\uff0c\u4e5f\u4e3a\u5176\u5728\u5546\u4e1a\u5e94\u7528\u4e2d\u7684\u63a8\u5e7f\u5960\u5b9a\u4e86\u57fa\u7840\u3002<\/p>\n<p>Gemini Embedding \u7684\u9ad8\u6027\u80fd\u4f7f\u5176\u5728\u4ee5\u4e0b\u9886\u57df\u5177\u6709\u5e7f\u6cdb\u7684\u5e94\u7528\u524d\u666f\uff1a<\/p>\n<ul>\n<li>\u641c\u7d22\u5f15\u64ce\uff1a\u63d0\u5347\u641c\u7d22\u7ed3\u679c\u7684\u76f8\u5173\u6027\uff0c\u652f\u6301\u8c37\u6b4c\u6b63\u5728\u6d4b\u8bd5\u7684\u7eaf AI \u9a71\u52a8\u641c\u7d22\u6a21\u5f0f\u3002<\/li>\n<li>\u591a\u8bed\u8a00\u5e94\u7528\uff1a\u589e\u5f3a\u8de8\u8bed\u8a00\u7ffb\u8bd1\u3001\u5ba2\u6237\u670d\u52a1\u81ea\u52a8\u5316\u548c\u5185\u5bb9\u6392\u540d\u80fd\u529b\u3002<\/li>\n<li>\u4f01\u4e1a\u670d\u52a1\uff1a\u4f18\u5316\u57fa\u4e8e Google Cloud \u7684 AI \u5206\u6790\u3001\u8bed\u4e49\u641c\u7d22\u548c\u81ea\u52a8\u5316\u6570\u636e\u68c0\u7d22\u529f\u80fd\u3002<\/li>\n<\/ul>\n","protected":false},"excerpt":{"rendered":"<p>On March 10, Google announced in a blog post on March 7 the launch of Gemini Embedding, an AI-based text processing model that is now integrated into the Gemini API. The model came out on top in the Massive Text Embedding Benchmark (MTEB), outperforming competitors such as Mistral, Cohere, and Qwen to become the most powerful text embedding model available today, Cohere, and Qwen as the most powerful text embedding model available. Gemini Embedding transforms text into numerical representations (vectors) to support semantic search, recommender systems, and document retrieval. It performs well in the MTEB benchmarks, with an average task<\/p>","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"footnotes":""},"categories":[146],"tags":[5932,559,5806,281],"collection":[],"class_list":["post-30365","post","type-post","status-publish","format-standard","hentry","category-news","tag-gemini-embedding","tag-mistral","tag-qwen","tag-281"],"acf":[],"_links":{"self":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/30365","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/comments?post=30365"}],"version-history":[{"count":0,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/30365\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/media?parent=30365"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/categories?post=30365"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/tags?post=30365"},{"taxonomy":"collection","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/collection?post=30365"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}