{"id":7479,"date":"2026-01-20T19:09:44","date_gmt":"2026-01-20T11:09:44","guid":{"rendered":"https:\/\/infernews.com\/?p=7479"},"modified":"2026-01-20T19:09:46","modified_gmt":"2026-01-20T11:09:46","slug":"glm-4-7-flash-%e5%9c%a8-mac-%e4%b8%8a%e7%9a%84%e6%b8%ac%e8%a9%a6%e5%8f%8a%e6%af%94%e8%bc%83","status":"publish","type":"post","link":"https:\/\/infernews.com\/blog\/glm-4-7-flash-%e5%9c%a8-mac-%e4%b8%8a%e7%9a%84%e6%b8%ac%e8%a9%a6%e5%8f%8a%e6%af%94%e8%bc%83\/","title":{"rendered":"GLM-4.7-Flash \u5728 Mac \u4e0a\u7684\u6e2c\u8a66\u53ca\u6bd4\u8f03"},"content":{"rendered":"\n<p><a href=\"https:\/\/huggingface.co\/zai-org\/GLM-4.7-Flash\" target=\"_blank\" rel=\"noreferrer noopener\">GLM-4.7-Flash<\/a> \u662f Zhipu AI \u6700\u65b0\u767c\u5e03\u7684 30B \u53c3\u6578 MoE \u6a21\u578b\uff083B \u6d3b\u8e8d\u53c3\u6578\uff09\uff0c\u5c08\u70ba\u9ad8\u6548\u672c\u5730\u904b\u884c\u8207\u7a0b\u5f0f\u78bc\u751f\u6210\u8a2d\u8a08\uff0c\u5728\u540c\u5c3a\u5bf8\u6a21\u578b\u4e2d\u9054\u5230\u958b\u6e90 SOTA \u6548\u80fd\u3002\u200b\u200b<\/p>\n\n\n\n<p>\u5f71\u7247\u4f7f\u7528 Inferencer app \u5728 M3 Ultra Mac Studio (512GB RAM) \u6e2c\u8a66 GLM-4.7-Flash \u7684 MLX \u91cf\u5316\u7248\u672c\uff0c\u6bd4\u8f03\u672a\u91cf\u5316\u8207 Q4\/Q5\/Q6\/Q8 \u6548\u80fd\u3002\u672a\u91cf\u5316\u7248\u751f\u6210 5000 \u500b token \u7684 3D \u592a\u967d\u7cfb\u7a0b\u5f0f\uff08\u542b\u6ed1\u9f20\u4e92\u52d5\uff09\uff0c\u512a\u65bc Qwen3-Coder 30B (1700 token) \u8207 Neotron\u3002<\/p>\n\n\n\n<p><a href=\"https:\/\/docs.z.ai\/guides\/llm\/glm-4.7\" target=\"_blank\" rel=\"noreferrer noopener\"><\/a>\u200b\u91cf\u5316\u5f8c Q5\/Q6 \u7248\u7dad\u6301\u9ad8\u54c1\u8cea\u8f38\u51fa\uff0856 token\/s\uff0c24-27GB \u8a18\u61b6\u9ad4\uff09\uff0c\u9069\u5408 32GB \u7cfb\u7d71\uff1b\u6279\u6b21\u8655\u7406 4 \u500b\u63d0\u793a\u9054 120 token\/s \u7e3d\u541e\u5410\u91cf\uff0c\u4f46\u8a18\u61b6\u9ad4\u5347\u81f3 140GB\u3002\u200b\u91cf\u5316\u6307\u6a19\u986f\u793a Q6 perplexity 1.23\u3001token accuracy 96.65%\uff0c\u50c5\u8f15\u5fae\u767c\u6563\uff0c\u8b49\u660e\u54c1\u8cea\u63a5\u8fd1\u57fa\u6a21\u3002<\/p>\n\n\n<figure class=\"wp-block-embed-youtube wp-block-embed is-type-video is-provider-youtube wp-embed-aspect-16-9 wp-has-aspect-ratio\"><div class=\"lyte-wrapper\" title=\"Let&amp;#039;s Run GLM-4-7-Flash - Local AI Super-Intelligence for the Rest of Us | REVIEW\" style=\"width:853px;max-width:100%;margin:5px auto;\"><div class=\"lyMe\" id=\"WYL_O5PI868ApCI\" itemprop=\"video\" itemscope itemtype=\"https:\/\/schema.org\/VideoObject\"><div><meta itemprop=\"thumbnailUrl\" content=\"https:\/\/infernews.com\/blog\/wp-content\/plugins\/wp-youtube-lyte\/lyteCache.php?origThumbUrl=https%3A%2F%2Fi.ytimg.com%2Fvi%2FO5PI868ApCI%2Fhqdefault.jpg\" \/><meta itemprop=\"embedURL\" content=\"https:\/\/www.youtube.com\/embed\/O5PI868ApCI\" \/><meta itemprop=\"duration\" content=\"PT22M40S\" \/><meta itemprop=\"uploadDate\" content=\"2026-01-20T06:14:55Z\" \/><\/div><div id=\"lyte_O5PI868ApCI\" data-src=\"https:\/\/infernews.com\/blog\/wp-content\/plugins\/wp-youtube-lyte\/lyteCache.php?origThumbUrl=https%3A%2F%2Fi.ytimg.com%2Fvi%2FO5PI868ApCI%2Fhqdefault.jpg\" class=\"pL\"><div class=\"tC\"><div class=\"tT\" itemprop=\"name\">Let&#039;s Run GLM-4-7-Flash - Local AI Super-Intelligence for the Rest of Us | REVIEW<\/div><\/div><div class=\"play\"><\/div><div class=\"ctrl\"><div class=\"Lctrl\"><\/div><div class=\"Rctrl\"><\/div><\/div><\/div><noscript><a href=\"https:\/\/youtu.be\/O5PI868ApCI\" rel=\"nofollow\"><img loading=\"lazy\" decoding=\"async\" src=\"https:\/\/infernews.com\/blog\/wp-content\/plugins\/wp-youtube-lyte\/lyteCache.php?origThumbUrl=https%3A%2F%2Fi.ytimg.com%2Fvi%2FO5PI868ApCI%2F0.jpg\" alt=\"Let&amp;#039;s Run GLM-4-7-Flash - Local AI Super-Intelligence for the Rest of Us | REVIEW\" width=\"853\" height=\"460\" \/><br \/>Watch this video on YouTube<\/a><\/noscript><meta itemprop=\"description\" content=\"Zhipu have released another benchmark topping for its size update to their GLM model. So let&#039;s see how well it performs locally. NOTE Thinking was disabled for these tests, watch the thinking version here: https:\/\/youtu.be\/2s292axPhwE TEST SYSTEM Inferencer App v1.9.3: https:\/\/inferencer.com 2025 M3 Ultra Mac Studio | 512GB RAM Q5: https:\/\/huggingface.co\/inferencerlabs\/GLM-4.7-Flash-MLX-5.5bit Q6: https:\/\/huggingface.co\/inferencerlabs\/GLM-4.7-Flash-MLX-6.5bit BUY NOW Mac Studio: https:\/\/vtudio.com\/a\/?a=mac+studio MacBook Pro: https:\/\/vtudio.com\/a\/?a=macbook+pro LG C2 42&quot; Monitor: https:\/\/vtudio.com\/a\/?a=lg+c2+42 Recommended NAS Drive: https:\/\/vtudio.com\/a\/?a=qnap+tvs-872xt COMPANION VIDEOS GLM 4.7: https:\/\/youtu.be\/E-8KJpUFalM Kimi K2 Thinking: https:\/\/youtu.be\/y6U36dO2jk0 Z-Image-Turbo: https:\/\/youtu.be\/RG5aSqRxAws Mac Studio Review: https:\/\/youtu.be\/-3ewAcnuN30 SPECIAL THANKS Thanks for your support and if you have any suggestions or would like to help us produce more videos, please visit: https:\/\/vtudio.com\/a\/?support Links to products often include an affiliate tracking code which allow us to earn fees on purchases you make through them.\"><\/div><\/div><div class=\"lL\" style=\"max-width:100%;width:853px;margin:5px auto;\"><\/div><figcaption><\/figcaption><\/figure>\n\n\n<figure class=\"wp-block-table\"><table class=\"has-fixed-layout\"><thead><tr><th class=\"has-text-align-left\" data-align=\"left\">\u91cf\u5316\u7d1a\u5225<\/th><th class=\"has-text-align-left\" data-align=\"left\">Perplexity<\/th><th class=\"has-text-align-left\" data-align=\"left\">Token Accuracy<\/th><th class=\"has-text-align-left\" data-align=\"left\">\u8a18\u61b6\u9ad4\u4f7f\u7528 (GB)<\/th><th class=\"has-text-align-left\" data-align=\"left\">Token\/s (\u55ae\u4e00\u6279\u6b21)<\/th><\/tr><\/thead><tbody><tr><td class=\"has-text-align-left\" data-align=\"left\">Base<\/td><td class=\"has-text-align-left\" data-align=\"left\">1.22<\/td><td class=\"has-text-align-left\" data-align=\"left\">100%<\/td><td class=\"has-text-align-left\" data-align=\"left\">60<\/td><td class=\"has-text-align-left\" data-align=\"left\">&#8211;<\/td><\/tr><tr><td class=\"has-text-align-left\" data-align=\"left\">Q5.5<\/td><td class=\"has-text-align-left\" data-align=\"left\">1.25<\/td><td class=\"has-text-align-left\" data-align=\"left\">94.5%<\/td><td class=\"has-text-align-left\" data-align=\"left\">24<\/td><td class=\"has-text-align-left\" data-align=\"left\">56<\/td><\/tr><tr><td class=\"has-text-align-left\" data-align=\"left\">Q6.5<\/td><td class=\"has-text-align-left\" data-align=\"left\">1.23<\/td><td class=\"has-text-align-left\" data-align=\"left\">96.7%<\/td><td class=\"has-text-align-left\" data-align=\"left\">27<\/td><td class=\"has-text-align-left\" data-align=\"left\">56<\/td><\/tr><tr><td class=\"has-text-align-left\" data-align=\"left\">Q8.5<\/td><td class=\"has-text-align-left\" data-align=\"left\">1.23<\/td><td class=\"has-text-align-left\" data-align=\"left\">97.8%<\/td><td class=\"has-text-align-left\" data-align=\"left\">34<\/td><td class=\"has-text-align-left\" data-align=\"left\">50<\/td><\/tr><\/tbody><\/table><\/figure>\n","protected":false},"excerpt":{"rendered":"<p>GLM-4.7-Flash \u662f Zhipu AI \u6700\u65b0\u767c\u5e03\u7684 30B \u53c3\u6578 MoE \u6a21\u578b\uff083B \u6d3b\u8e8d\u53c3\u6578\uff09\uff0c\u5c08\u70ba\u9ad8\u6548\u672c\u5730\u904b\u884c\u8207\u7a0b\u5f0f\u78bc\u751f\u6210\u8a2d\u8a08\uff0c\u5728\u540c\u5c3a\u5bf8\u6a21\u578b\u4e2d\u9054\u5230\u958b\u6e90 SOTA \u6548\u80fd\u3002\u200b\u200b \u5f71\u7247\u4f7f\u7528 Inferencer app \u5728 M3 Ultra Mac Studio (512GB RAM) \u6e2c\u8a66 GLM-4.7-Flash \u7684 MLX \u91cf\u5316\u7248\u672c\uff0c\u6bd4\u8f03\u672a\u91cf\u5316\u8207 Q4\/Q5\/Q6\/Q8 \u6548\u80fd\u3002\u672a\u91cf\u5316\u7248\u751f\u6210 5000 \u500b token \u7684 3D \u592a\u967d\u7cfb\u7a0b\u5f0f\uff08\u542b\u6ed1\u9f20\u4e92\u52d5\uff09\uff0c\u512a\u65bc Qwen3-Coder 30B (1700 token) \u8207 Neotron\u3002 \u200b\u91cf\u5316\u5f8c Q5\/Q6 \u7248\u7dad\u6301\u9ad8\u54c1\u8cea\u8f38\u51fa\uff0856 token\/s\uff0c24-27GB \u8a18\u61b6\u9ad4\uff09\uff0c\u9069\u5408 32GB \u7cfb\u7d71\uff1b\u6279\u6b21\u8655\u7406 4 \u500b\u63d0\u793a\u9054 120 token\/s \u7e3d\u541e\u5410\u91cf\uff0c\u4f46\u8a18\u61b6\u9ad4\u5347\u81f3 140GB\u3002\u200b\u91cf\u5316\u6307\u6a19\u986f\u793a Q6 [&hellip;]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"googlesitekit_rrm_CAowvqSiDA:productID":"","footnotes":""},"categories":[76,133],"tags":[],"class_list":["post-7479","post","type-post","status-publish","format-standard","hentry","category-76","category-133"],"_links":{"self":[{"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/posts\/7479","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/comments?post=7479"}],"version-history":[{"count":1,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/posts\/7479\/revisions"}],"predecessor-version":[{"id":7480,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/posts\/7479\/revisions\/7480"}],"wp:attachment":[{"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/media?parent=7479"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/categories?post=7479"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/tags?post=7479"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}