[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"news-slug-qwen3-8-max-open-weights-stripped-relicense":3,"news-related-b863d01a-dfdf-41c9-8266-e4602e58bde3":38},{"id":4,"title":5,"summary":6,"content":7,"original_url":8,"source_id":9,"tags":10,"translations":24,"news_slug":31,"published_at":32,"created_at":33,"modified_at":34,"is_published":35,"publish_type":36,"image_url":14,"view_count":37},"b863d01a-dfdf-41c9-8266-e4602e58bde3","Qwen3.8-Max 开源权重落地:砍掉视觉与 1M 上下文,许可证换成收入分成","阿里 Qwen3.8-Max 开放权重 8 月 12 日上线 Hugging Face:2.4T 参数 MoE 可下载,但砍掉视觉与 1M 上下文、强制思考模式,并弃用 Apache 2.0 改用带收入分成的新许可证,社区评价转负。","8 月 12 日,阿里把 Qwen3.8-Max 的开放权重放上了 Hugging Face:两个官方仓库 Qwen\u002FQwen3.8-2.4T-A95B 与对应的 FP8 量化版,2.4 万亿总参数、950 亿激活的细粒度 MoE,训练后(post-trained)检查点,官方文档明确面向 vLLM、SGLang、TokenSpeed 部署路径。NVIDIA 同日发布的部署博客也确认了这次开放。这是 Qwen 家族第一次把 Max 级旗舰的权重真正交出去——但下载下来你会发现,它不是你在 API 里试过的那个模型。\n\n## 开出来的不是旗舰,是旗舰的文本子集\n\n对照托管版 Qwen3.8-Max,开放权重版砍掉了三样东西:视觉输入(API 版有,开源版纯文本)、最高 1M token 的上下文窗口(开源版原生窗口明显更小)、可选的思考模式(开源版强制所有交互走 thinking)。工具调用和 Qoder 那套智能体能力也不在基座检查点里。\n\n这个落差直接点燃了社区。Hugging Face 仓库的讨论区在权重上线几小时内就出现了长帖,有评论认为砍掉视觉「移除了它一半的核心价值」,还有人把这种「同名不同能力」的发布模式比作游戏行业的 DLC 分段付费,称这一操作耗尽了 Qwen 前几代开源积累的信任。\n\n硬件门槛也把绝大多数人挡在门外。BF16 全量权重约 4.89TB;Unsloth 提供的 GGUF 量化从 1-bit 的约 397GB 到 4-bit 的约 1.31TB——就算压到极限,依然是多卡工作站或重型 NVMe offload 的级别。NVIDIA 的参考部署是 GB300 NVL72(72 块 Blackwell Ultra GPU),FP8 下单卡吞吐超过 4000 token\u002Fs。对单卡用户来说,这次开放基本与本地部署无关。\n\n## 许可证才是真正的信号\n\n比减配更值得注意的,是许可证换了。此前 Qwen 各代普遍采用宽松的 Apache 2.0,这次 HF 仓库挂的是一份自定义的 qwen3.8-max 许可证,围绕发布的报道描述其对大型商业用户有收入分成条款,具体阈值和比例还在敲定中。Hugging Face 8 月 14 日发布的 Summer 2026 开源模型报告也印证了这个方向:报告指出行业最前沿正在转向更明确的商业化路径,Kimi K3 与 Qwen 3.8 Max 近期都加入了非商业限制与收入分成要求。\n\n这解释了「为什么砍」。当开源权重从获客手段变成需要直接变现的资产,旗舰能力就必须留在 API 侧。阿里仍然贡献了 Qwen 家族史上第一次 Max 级开放,但在「开放」的定义上,这次释放更像一场研究基础设施的展示,而不是对社区的完整交接。\n\n## 27B 补上了,但问题的答案没变\n\n8 月 15 日,承诺中面向单卡的 Qwen3.8-27B 也已上线(上线当日冲上 Hacker News 首页),Apache 2.0,这次没再缩水。这让「缺的那一半」有了着落——单卡用户有了可自托管的选择。\n\n但对整个行业来说,真正的问题没变:当顶级模型的开放从「全量交付」变成「分层交付」,开源社区拿到了权重,却没有拿到与闭源版对等的能力。你在 Qoder 里测的是完整版,下载到本地的是阉割版——以后评估任何「开源旗舰」,都得先问一句:开的是哪一半?参考:[explainx.ai 分析](https:\u002F\u002Fwww.explainx.ai\u002Fblog\u002Fqwen3-8-max-open-weights-live-hugging-face-august-2026)","https:\u002F\u002Fwww.explainx.ai\u002Fblog\u002Fqwen3-8-max-open-weights-live-hugging-face-august-2026","8cb75837-7ecc-4e8a-bc52-167b41f1be2f",[11,15,18,21],{"id":12,"name":13,"slug":13,"description":14,"color":14},"a8002d98-9df1-4ab9-94d4-a7625af634c4","china-ai",null,{"id":16,"name":17,"slug":17,"description":14,"color":14},"8ddf2b28-0234-41a4-9862-3f0faef96472","market-analysis",{"id":19,"name":20,"slug":20,"description":14,"color":14},"b9bd9039-fcdb-41a8-b85b-fc1587def2b9","open-source",{"id":22,"name":23,"slug":23,"description":14,"color":14},"c187600e-804c-4697-b828-1e4330e0eb10","qwen",[25],{"id":26,"lang":27,"title":28,"summary":29,"content":30},"6a7ec18c-5038-4815-83d9-cf8a97bd2c97","en","Qwen3.8-Max Open Weights Arrive Stripped, Under a New License","Alibaba shipped Qwen3.8-Max open weights to Hugging Face on August 12: a downloadable 2.4T-parameter MoE, but with vision and 1M context removed, thinking mode forced on, and a new revenue-share license replacing Apache 2.0. Community reaction turned negative.","On August 12, Alibaba published the open weights of Qwen3.8-Max on Hugging Face: two official repositories, Qwen\u002FQwen3.8-2.4T-A95B and an FP8 quantized variant, a fine-grained Mixture-of-Experts with 2.4 trillion total parameters and 95 billion active. It is a post-trained checkpoint, with the model card explicitly targeting vLLM, SGLang, and TokenSpeed deployment paths. NVIDIA confirmed the release the same day in a deployment engineering blog. This is the first time the Qwen family has handed out a Max-tier flagship — but once you download it, you will find it is not the model you tried in the API.\n\n## What landed is not the flagship, but a text-only subset\n\nCompared against the hosted Qwen3.8-Max, the open-weights release is missing three things: vision input (available on the API, absent here — the checkpoint is text-only), the up-to-1M-token context window (the native window is documented well below that), and optional thinking mode (the open version forces thinking on for all interactions). Tool use and the Qoder agent surface are also not part of the base checkpoint.\n\nThe gap immediately lit up the community. A Hugging Face discussion thread opened within hours of the drop: commenters argued that stripping vision \"removes half its core value,\" others compared the same-name-different-capabilities release pattern to game-industry DLC paywalling, with at least one saying it cost Alibaba \"all good will\" built up across earlier open Qwen generations.\n\nHardware requirements keep most people out anyway. The full BF16 checkpoint weighs around 4.89 TB; Unsloth GGUF quants span roughly 397 GB at 1-bit to about 1.31 TB at 4-bit — even pushed to the floor, that is multi-GPU workstation or heavy NVMe-offload territory. NVIDIA reference deployment is a GB300 NVL72 rack (72 Blackwell Ultra GPUs), serving the FP8 checkpoint at over 4,000 tokens\u002Fsec per GPU. For single-card users, this release has essentially nothing to do with local deployment.\n\n## The license is the real signal\n\nMore telling than the feature cuts is the license change. Earlier Qwen generations shipped under permissive Apache 2.0 terms; this time the HF repo carries a bespoke qwen3.8-max license, and reporting around the launch describes a revenue-sharing requirement for large commercial users, with the exact threshold and percentage still being finalized. Hugging Face Summer 2026 State of Open Models report, published August 14, corroborates the direction: the very top of the frontier is beginning to shift toward clearer monetization, with Kimi K3 and Qwen 3.8 Max recently adding non-commercial restrictions and revenue-share requirements.\n\nThat explains the \"why\" behind the cuts. When open weights shift from a customer-acquisition channel into an asset that must monetize directly, flagship capabilities have to stay on the API side. Alibaba still delivered the first Max-tier open release in Qwen history — but in the definition of \"open,\" this drop reads more like a research-infrastructure showcase than a full handoff to the community.\n\n## The 27B shipped, but the real question remains\n\nOn August 15, the promised single-GPU companion Qwen3.8-27B also went live (it hit #1 on Hacker News the day it shipped), under Apache 2.0, with no feature stripping this time. That gives the \"missing half\" a landing spot — self-hosters now have a genuinely accessible option.\n\nBut for the industry at large, the underlying question is unchanged: when top-tier openness turns from full delivery into tiered delivery, the open-source community gets the weights without getting capabilities on par with the closed version. What you tested in Qoder is the full model; what you download locally is the stripped one. From now on, evaluating any \"open flagship\" starts with one question: which half is actually open? Reference: [explainx.ai analysis](https:\u002F\u002Fwww.explainx.ai\u002Fblog\u002Fqwen3-8-max-open-weights-live-hugging-face-august-2026)","qwen3-8-max-open-weights-stripped-relicense","2026-08-23T13:30:00Z","2026-08-22T21:13:31.501182Z","2026-08-22T21:13:31.501190Z",true,"agent",55,{"items":39},[40,45,50,55,60,65],{"id":41,"title":42,"news_slug":43,"published_at":44},"0d8fdf45-4585-47c0-9e78-3652e318b156","Apple Intelligence 中国版落地:通义千问接管语言 AI,百度负责视觉搜索","apple-intelligence-china-qwen-baidu-2026","2026-08-25T12:00:00+00:00",{"id":46,"title":47,"news_slug":48,"published_at":49},"1844afb1-3a1c-4acd-9e4c-f5e2792a2018","下载免费不等于商用免费：HF Summer 2026 隐藏的开源前沿许可证分水岭","frontier-license-shift-hf-summer-2026","2026-08-23T12:30:00+00:00",{"id":51,"title":52,"news_slug":53,"published_at":54},"8d7b30e0-996f-4141-8501-8f464bda6282","中美开放权重参数上限差距拉到 20 倍:Hugging Face 夏季报告里的三条隐藏数据","hugging-face-summer-2026-frontier-ceiling","2026-08-22T14:00:00+00:00",{"id":56,"title":57,"news_slug":58,"published_at":59},"4bb31ede-b9c4-4762-86ae-9d3b008557ca","Hugging Face Summer 2026 报告:Qwen 拿下 15 万衍生模型, GGUF 仓库一年涨 464%","hugging-face-state-of-open-models-summer-2026","2026-08-18T02:00:00+00:00",{"id":61,"title":62,"news_slug":63,"published_at":64},"aa4d3e55-383d-4855-9965-cc6a4d2e38a7","Qwen 下载量 6 个月破 30 亿:开源模型的「默认底座」第一次换成了中国厂商","qwen-3-billion-downloads-open-weights","2026-08-15T23:20:00+00:00",{"id":66,"title":67,"news_slug":68,"published_at":69},"ad3be632-49a1-44f5-9816-62c168e56467","全球大模型调用量榜前五全是\"中国造\":开源 MoE 正在重写 OpenRouter 的地理坐标","openrouter-top5-china-moe-open-source-2026w31","2026-08-02T03:30:00+00:00"]