[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"article-glm-5-2-tops-open-rankings-huawei-chips-en":3,"article-related-glm-5-2-tops-open-rankings-huawei-chips-en":30,"series-industry-3626d9c1-ebb7-431f-9b9d-80f3bdaaa976":75},{"id":4,"slug":5,"title":6,"content":7,"summary":8,"source":9,"source_url":10,"author":11,"image_url":12,"cover_image":12,"category":13,"language":14,"translated_content":11,"related_article_id":15,"keywords":16,"key_takeaways":22,"views":26,"created_at":27,"published_at":28,"topic_cluster_id":29},"3626d9c1-ebb7-431f-9b9d-80f3bdaaa976","glm-5-2-tops-open-rankings-huawei-chips-en","GLM-5.2 tops open AI rankings on Huawei chips","\u003Cp data-speakable=\"summary\">Z.ai’s GLM-5.2 is the top open-weight AI model on Huawei chips after \u003Ca href=\"\u002Ftag\u002Fanthropic\">Anthropic\u003C\u002Fa>’s Fable 5 ban.\u003C\u002Fp>\u003Cp>\u003Ca href=\"https:\u002F\u002Fwww.techtimes.com\u002Farticles\u002F318810\u002F20260621\u002Fchina-ai-parity-glm-52-tops-open-rankings-on-huawei-chips-as-fable-5-stays-banned.htm\" target=\"_blank\" rel=\"noopener\">Techtimes\u003C\u002Fa> reports that Z.ai’s GLM-5.2 now leads open rankings after a June 17 release, while \u003Ca href=\"https:\u002F\u002Fwww.anthropic.com\" target=\"_blank\" rel=\"noopener\">Anthropic\u003C\u002Fa> keeps Fable 5 and Mythos 5 offline following a US export order. The model was trained entirely on \u003Ca href=\"https:\u002F\u002Fwww.huawei.com\" target=\"_blank\" rel=\"noopener\">Huawei\u003C\u002Fa> Ascend 910B chips, with no \u003Ca href=\"\u002Ftag\u002Fnvidia\">Nvidia\u003C\u002Fa> hardware in the training stack.\u003C\u002Fp>\u003Ctable>\u003Cthead>\u003Ctr>\u003Cth>項目\u003C\u002Fth>\u003Cth>數值\u003C\u002Fth>\u003C\u002Ftr>\u003C\u002Fthead>\u003Ctbody>\u003Ctr>\u003Ctd>GLM-5.2 release date\u003C\u002Ftd>\u003Ctd>June 17, 2026\u003C\u002Ftd>\u003C\u002Ftr>\u003Ctr>\u003Ctd>Fable 5 ban date\u003C\u002Ftd>\u003Ctd>June 12, 2026\u003C\u002Ftd>\u003C\u002Ftr>\u003Ctr>\u003Ctd>Training chips\u003C\u002Ftd>\u003Ctd>100,000 Huawei Ascend 910B\u003C\u002Ftd>\u003C\u002Ftr>\u003Ctr>\u003Ctd>Total parameters\u003C\u002Ftd>\u003Ctd>744 billion\u003C\u002Ftd>\u003C\u002Ftr>\u003Ctr>\u003Ctd>Active parameters per inference\u003C\u002Ftd>\u003Ctd>~40 billion\u003C\u002Ftd>\u003C\u002Ftr>\u003Ctr>\u003Ctd>Context window\u003C\u002Ftd>\u003Ctd>1 million tokens\u003C\u002Ftd>\u003C\u002Ftr>\u003Ctr>\u003Ctd>Training cost estimate\u003C\u002Ftd>\u003Ctd>$25 million\u003C\u002Ftd>\u003C\u002Ftr>\u003C\u002Ftbody>\u003C\u002Ftable>\u003Ch2>What changed\u003C\u002Fh2>\u003Cp>Z.ai says GLM-5.2 is the first model family it trained entirely on domestic Chinese silicon, and the release landed with MIT-licensed weights on \u003Ca href=\"https:\u002F\u002Fhuggingface.co\" target=\"_blank\" rel=\"noopener\">Hugging Face\u003C\u002Fa>. That made the model easy to download, self-host, and test against closed Western systems.\u003C\u002Fp>\n\u003Cfigure class=\"my-6\">\u003Cimg src=\"https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1782635577270-n94l.png\" alt=\"GLM-5.2 tops open AI rankings on Huawei chips\" class=\"rounded-xl w-full\" loading=\"lazy\" \u002F>\u003C\u002Ffigure>\n\u003Cp>On public benchmarks, GLM-5.2 posted strong results in coding and design tasks. It ranked second on Code Arena with an Elo score of 1,595, led \u003Ca href=\"\u002Ftag\u002Fswe-bench\">SWE-bench\u003C\u002Fa> Pro with 62.1, and took first on Design Arena. The same report notes a wide spread on harder tests: GLM-5.2 scored 13.0 on SWE-Marathon, while \u003Ca href=\"\u002Ftag\u002Fclaude\">Claude\u003C\u002Fa> Opus 4.8 reached 26.0.\u003C\u002Fp>\u003Cul>\u003Cli>GLM-5.2 uses a Mixture-of-Experts design with 744 billion total parameters.\u003C\u002Fli>\u003Cli>About 40 billion parameters activate per inference.\u003C\u002Fli>\u003Cli>DeepSeek Sparse Attention helps make the 1M-token context window usable.\u003C\u002Fli>\u003Cli>Inference speed is about 17 to 19 tokens per second, slower than Nvidia-backed rivals.\u003C\u002Fli>\u003C\u002Ful>\u003Cp>The model’s architecture is built for long-context work, especially large codebases and agentic software tasks. Z.ai says the routing system picks 8 of 256 expert sub-networks for each \u003Ca href=\"\u002Ftag\u002Ftoken\">token\u003C\u002Fa>, which keeps compute lower than a dense model at similar scale.\u003C\u002Fp>\u003Ch2>Why it matters\u003C\u002Fh2>\u003Cp>The timing matters as much as the benchmarks. On June 12, the US Commerce Department ordered Anthropic to disable Fable 5 and Mythos 5 for foreign nationals worldwide, citing a jailbreak issue. Five days later, Z.ai shipped an open-weight alternative that developers can run locally and cannot be pulled back with an export order.\u003C\u002Fp>\n\u003Cfigure class=\"my-6\">\u003Cimg src=\"https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1782635569768-9i3j.png\" alt=\"GLM-5.2 tops open AI rankings on Huawei chips\" class=\"rounded-xl w-full\" loading=\"lazy\" \u002F>\u003C\u002Ffigure>\n\u003Cp>That shifts the policy debate from chip denial to practical access. If a Chinese model can match key commercial use cases on coding and design while running on Huawei hardware, export controls may slow training, but they do not stop frontier-class releases from reaching developers outside the US.\u003C\u002Fp>\u003Cp>There is still a gap on the hardest reasoning tests, and the report cites Epoch AI and Stanford’s 2026 AI Index to show that US labs remain ahead on evaluations built to resist gaming. But for teams choosing between closed APIs and local deployment, GLM-5.2 changes the default answer: open weights are now close enough to matter.\u003C\u002Fp>\u003Cp>The real question is no longer whether Chinese labs can ship competitive code models on domestic chips, but how much leverage export controls still have once those weights are public.\u003C\u002Fp>","Z.ai’s GLM-5.2 hit the top open-weight rank on Huawei Ascend chips, as Anthropic’s Fable 5 stayed offline after a US export ban.","www.techtimes.com","https:\u002F\u002Fwww.techtimes.com\u002Farticles\u002F318810\u002F20260621\u002Fchina-ai-parity-glm-52-tops-open-rankings-huawei-chips-fable-5-stays-banned.htm",null,"https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1782635577270-n94l.png","industry","en","8559088c-4d91-4b2e-9b61-840e92378880",[17,18,19,20,21],"GLM-5.2","Huawei Ascend","Anthropic","open-weight AI","export controls",[23,24,25],"GLM-5.2 is now the top open-weight model in the report’s rankings.","It was trained on 100,000 Huawei Ascend 910B chips with no Nvidia hardware.","US export controls slowed access to Fable 5, but did not block a rival open release.",0,"2026-06-28T08:32:29.011648+00:00","2026-06-28T08:32:29.003+00:00","520d4d46-4f94-4abd-a36f-af3a6c940a98",{"tags":31,"relatedLang":34,"relatedPosts":38},[32],{"name":19,"slug":33},"anthropic",{"id":15,"slug":35,"title":36,"language":37},"glm-5-2-tops-open-rankings-huawei-chips-zh","GLM-5.2 登頂華為晶片開源榜","zh",[39,45,51,57,63,69],{"id":40,"slug":41,"title":42,"cover_image":43,"image_url":43,"created_at":44,"category":13},"1334df83-1b4b-4c5f-9707-e020fe64f521","openclaw-openai-realtime-paid-api-not-subscription-perk-en","OpenClaw should treat OpenAI Realtime as a paid API, not a subscripti…","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1782674272037-c6qm.png","2026-06-28T19:17:24.994296+00:00",{"id":46,"slug":47,"title":48,"cover_image":49,"image_url":49,"created_at":50,"category":13},"04cdbad0-ca9e-4794-b2fd-79125566e019","krea-2-two-second-image-generation-teams-en","Krea 2 brings 2-second image generation to teams","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1782673372571-q3g0.png","2026-06-28T19:02:22.91619+00:00",{"id":52,"slug":53,"title":54,"cover_image":55,"image_url":55,"created_at":56,"category":13},"6622fa0c-3619-4bc2-adf1-7e3813fd5174","us-model-curbs-security-deals-not-bans-en","US model curbs should be lifted through security deals, not blanket b…","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1782658965406-pysv.png","2026-06-28T15:02:20.414547+00:00",{"id":58,"slug":59,"title":60,"cover_image":61,"image_url":61,"created_at":62,"category":13},"042e3873-18ef-472c-9e34-e5e0e3995e1e","metas-moderation-shift-shows-where-ai-cuts-costs-en","Meta’s moderation shift shows where AI cuts costs","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1782654472664-zyid.png","2026-06-28T13:47:23.092521+00:00",{"id":64,"slug":65,"title":66,"cover_image":67,"image_url":67,"created_at":68,"category":13},"ed5f0914-640e-4fe1-af0c-ead6acc6403a","meta-replacing-moderators-with-ai-to-cut-costs-en","Meta is replacing moderators with AI to cut costs","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1782653573805-ro1p.png","2026-06-28T13:32:30.212298+00:00",{"id":70,"slug":71,"title":72,"cover_image":73,"image_url":73,"created_at":74,"category":13},"679d7344-8847-4cae-8b8f-01f6a065aba6","metas-ai-moderation-push-is-the-wrong-tradeoff-en","Meta’s AI moderation push is the wrong tradeoff","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1782652668263-q22a.png","2026-06-28T13:17:22.301423+00:00",[76,81,86,91,96,101,106,111,116,121],{"id":77,"slug":78,"title":79,"created_at":80},"d35a1bd9-e709-412e-a2df-392df1dc572a","ai-impact-2026-developments-market-en","AI's Impact in 2026: Key Developments and Market Shifts","2026-03-25T16:20:33.205823+00:00",{"id":82,"slug":83,"title":84,"created_at":85},"5ed27921-5fd6-492e-8c59-78393bf37710","trumps-ai-legislative-framework-en","Trump's AI Legislative Framework: What's Inside?","2026-03-25T16:22:20.005325+00:00",{"id":87,"slug":88,"title":89,"created_at":90},"e454a642-f03c-4794-b185-5f651aebbaca","nvidia-gtc-2026-key-highlights-innovations-en","NVIDIA GTC 2026: Key Highlights and Innovations","2026-03-25T16:22:47.882615+00:00",{"id":92,"slug":93,"title":94,"created_at":95},"0ebb5b16-774a-4922-945d-5f2ce1df5a6d","claude-usage-diversifies-learning-curves-en","Claude Usage Diversifies, Learning Curves Emerge","2026-03-25T16:25:50.770376+00:00",{"id":97,"slug":98,"title":99,"created_at":100},"69934e86-2fc5-4280-8223-7b917a48ace8","openclaw-ai-commoditization-concerns-en","OpenClaw's Rise Raises Concerns of AI Model Commoditization","2026-03-25T16:26:30.582047+00:00",{"id":102,"slug":103,"title":104,"created_at":105},"b4b2575b-2ac8-46b2-b90e-ab1d7c060797","google-gemini-ai-rollout-2026-en","Google's Gemini AI Rollout Extended to 2026","2026-03-25T16:28:14.808842+00:00",{"id":107,"slug":108,"title":109,"created_at":110},"6e18bc65-42ae-4ad0-b564-67d7f66b979e","meta-llama4-fabricated-results-scandal-en","Meta's Llama 4 Scandal: Fabricated AI Test Results Unveiled","2026-03-25T16:29:15.482836+00:00",{"id":112,"slug":113,"title":114,"created_at":115},"bf888e9d-08be-4f47-996c-7b24b5ab3500","accenture-mistral-ai-deployment-en","Accenture and Mistral AI Team Up for AI Deployment","2026-03-25T16:31:01.894655+00:00",{"id":117,"slug":118,"title":119,"created_at":120},"5382b536-fad2-49c6-ac85-9eb2bae49f35","mistral-ai-high-stakes-2026-en","Mistral AI: Facing High Stakes in 2026","2026-03-25T16:31:39.941974+00:00",{"id":122,"slug":123,"title":124,"created_at":125},"9da3d2d6-b669-4971-ba1d-17fdb3548ed5","cursors-meteoric-rise-pressures-en","Cursor's Meteoric Rise Faces Industry Pressures","2026-03-25T16:32:21.899217+00:00"]