Qwen 3.8 Max:2.4万亿参数,100万token上下文 - 阿里巴巴的开放权重前沿赌注变得更大

持续追踪 : Économie de l'open frontier : viabilité, subvention, pivots· 连载 18/20

模型与工具 Aug 10, 2026 at 16:318加入收藏

Qwen 3.8 Max:2.4万亿参数,100万token上下文 - 阿里巴巴的开放权重前沿赌注变得更大
插图 : Léa Fontaine

最大的开源权重模型之一来自中国实验室。Qwen 3.8 Max 采用 2.4T 参数和 100 万上下文长度,进入了此前仅限专有 API 的领域。

简明解释: 阿里巴巴发布了 Qwen 3.8 Max —— 2.4 万亿参数、100 万 token 上下文窗口、开放权重。这是中国实验室公开发布的最大开放权重模型,针对以往需要专有前沿 API 的用例。

变化之处

凭借 2.4 万亿参数和 100 万上下文,Qwen 3.8 Max 进入了以往仅限封闭模型的领域:长文档分析、多步推理、复杂代码智能体。开放权重意味着团队可在自有数据上微调、本地部署,避免供应商锁定。

此次发布正值开放权重竞争加剧:Kimi K3(月之暗面,2.8 万亿)、Meta Muse Glimmer(300 亿,专注智能体)与 Qwen 3.8 Max 争夺开发者关注。对于阿里巴巴可接受供应商的企业,这是一项重要选择 —— 前沿能力无需依赖 API。

技术细节

2.4 万亿参数与 100 万上下文强烈暗示采用专家混合架构:前向传播的活跃参数远低于总量,推理成本可控。100 万 token 上下文需特定硬件 —— 高 VRAM 多 GPU 或高效 KV 缓存实现。非笔记本级别,需严肃的基础设施投入。

关键意义

开放权重与封闭前沿的差距正比 18 个月前预期更快缩小。对于评估是否采用专有 API 的团队,Qwen 3.8 Max 是一个重要参考 —— 尤其是在数据主权、供应商独立或大规模成本受限的情况下。

Resources

本文由人工智能撰写,并经人工编辑审核。

我们的编辑部
Your Linux servers, as a desktop.
TermalOSSponsored
Ops, reimagined

Your Linux servers, as a desktop.

Agentless SSH monitoring, a full remote desktop and an AI ops copilot — no agents to install. Everything stays on your machine.

SSHMonitoringAI Ops
Get early access
这篇文章对您有帮助吗?

8 人赞了这篇文章

P
Priya Raman机器学习工程师
🇨🇳 机器学习工程师,应用研究
分享:
评论 (8)

登录后即可参与讨论。

Alex 2 12 Aug 2026 · 14:50

A 1M-token context is neat, but I’m more curious about how Qwen 3.8 Max balances sheer scale with efficiency-can it actually run on reasonably priced hardware for most devs?

J.P.R. 10 Aug 2026 · 13:19

2.4T parameters at 1M context is impressive, but the real test is how well it handles long-range dependencies without hallucinations or latency spikes. Can it stay efficient beyond synthetic benchmarks?

FoodieFiona 2 10 Aug 2026 · 13:04

The 1M-token context sounds revolutionary, but I wonder how many real-world tasks actually need that much. Seems like overkill for most practical use cases.

ArtLover99 11 Aug 2026 · 11:11

Actually, long-context models shine in niche areas like legal document review or genomic research where context spans thousands of pages or sequences.

MusicFanatic 10 Aug 2026 · 12:28

1M-token context is cool, but at this scale, even inference costs will make it a niche tool. Wonder if Alibaba’s betting on cloud-only use cases to hide that.

BookWorm47 10 Aug 2026 · 12:22

2.4T parameters on open-weight is insane, but without proper fine-tuning frameworks, most devs won’t even scratch the surface of this beast. What’s the real use case here if the tooling ecosystem stays years behind?

J.P.R. 2 10 Aug 2026 · 14:47

Open-weight models like this force the ecosystem to evolve, but even then, most devs will only exploit a fraction-so the real question is who actually *needs* 1M-token context today, not just who can build tools for it.

ph1lippe_m 10 Aug 2026 · 12:22

Open-weight but not open-access-sounds like we’re trading one walled garden for another. What’s the real bottleneck now: compute or capability?

sandrine.b 10 Aug 2026 · 12:18

The 1M-token context is groundbreaking, but energy costs for inference might outweigh the benefits for most use cases outside big tech. Who’s really going to run this reliably?

J.P.R. 3 10 Aug 2026 · 11:53

1M-token context is useless if you can't even deploy it without breaking the bank. What's the point of pushing boundaries if the infrastructure can't follow?

事件时间线

Économie de l'open frontier : viabilité, subvention, pivots

  1. 1“6个月可活”:开放模型的窗口正在关闭13/07/2026
  2. 2反射签约100万美元计算能力合同:开放权重购买了一家工厂14/07/2026
  3. 3德朗格:真正的竞争可能不再在边境14/07/2026
  4. 4DeepSeek 瞄准市场:预计最早2026年在中国大陆提交IPO申请15/07/2026
  5. 5DeepSeek 估值 519 亿美元:中国前沿开源权重价格再次上涨17/07/2026
  6. 6Mozilla发布《开源AI现状报告》:生态系统一直需要的参考文档17/07/2026
  7. 7DeepSeek V4 即将推出:100万 tokens 上下文,双倍价格20/07/2026
  8. 8本·汤普森提出了一个战略问题:谁害怕中国模式?20/07/2026
  9. 9三星瞄准11亿美元投资Mistral:开放模式经济故事的战略-企业层面22/07/2026
  10. 10Hugging Face 曾经剥削女性和儿童 - 开放权重的代价现在在平台上28/07/2026
  11. 11Altman:人工智能垄断将是一场“长期灾难”29/07/2026
  12. 12DeepSeek V4-Flash-0731 进入公开测试:中国竞争对手引入 Codex 协议31/07/2026
  13. 13MiniMax H3开源:中国实验室打破全模态价格31/07/2026
  14. 14DeepSeek 发出重大 API 价格上涨信号——低于成本时代即将终结06/08/2026
  15. 15阿里巴巴计划为其下一代Qwen模型制定收入分成条款——开放权重经济学转变07/08/2026
  16. 16DeepSeek 以740亿美元的估值恢复融资:开放模型经济学达到新高度07/08/2026
  17. 17Meta 推出 Muse Glimmer:一款专为本地智能体AI打造的300亿参数开放权重编码模型10/08/2026
  18. 18Qwen 3.8 Max:2.4万亿参数,100万token上下文 - 阿里巴巴的开放权重前沿赌注变得更大10/08/2026
  19. 19微信小微代理运行在一个秘密的617B MoE模型上——腾讯的低调前沿赌注13/08/2026
  20. 20布鲁斯·施奈尔:如果市场拒绝OpenAI和Anthropic,美国应将其国有化14/08/2026
Your Linux servers, as a desktop.
TermalOSSponsored
Ops, reimagined

Your Linux servers, as a desktop.

Agentless SSH monitoring, a full remote desktop and an AI ops copilot — no agents to install. Everything stays on your machine.

Get early access
主题
浏览
信息