Qwen3.8-Flash-Next 与 GLM-5.3-Flash 两款开放权重模型发布，均为 MoE 架构，每 token 分别激活 6B 和 18B 参数。
推荐理由：推文列出两款开放权重模型的参数规模与多项基准结果，可供读者对比其与前沿闭源模型的差距。
开源生态
全部主题开源模型、框架与仓库动态：权重开放、社区项目爆火、开源与闭源的力量消长。
当前仅显示精选新闻357条精选相关主题模型发布Hugging FaceAI 编码
最新精选
第 141–160 条 · 共 357 条@kimmonismus@kimmonismus精选AI 评分7272 @omarsar0@omarsar0精选AI 评分7676 
引用@Alibaba_Qwen@Alibaba_Qwen⚡Meet Qwen3.8-Flash, a multimodal MoE and an early preview of the Qwen4 architecture, now open-weight! The production version Qwen3.8-Flash will be available soon via QwenCloud API at just $ 0.16/1M input tokens and $ 0.47/1M output tokens. 125B parameters + 51B N-gram embeddings, with just 6B activated per token. Unmatched cost-efficiency. What's new: 🥳 - Next architecture: GDN + QSA hybrid attention, Gated Residual, N-gram Embedding & Muon optimizer, serving as a precursor to the architecture used in Qwen4. - Dramatically lower training and inference costs: trained at just 1/9 the cost of Qwen3.7-Plus, while outperforming it across the board with especially strong gains in coding and office tasks. - Strong performance: scoring 58.7 on DeepSWE 1.1, 62.5 on SWE-bench Pro, 73.9 on CoWorkBench, 84.5 on AndroidWorld, and 95.7 on MathVision (with CI). - 262K native context, extensible to 1M with YaRN. We’re also releasing the weights for Qwen3.8-Flash-Next, giving the community an early look at the new architecture we’re exploring for Qwen4.🚀 We can't wait to see what you build with Qwen3.8-Flash!👀👇 - Blog: https://t.co/M5hYypFLgJ - Technical Report: https://t.co/IF0gObIkQO - Hugging Face: https://t.co/6ow8QVAABt - ModelScope: https://t.co/tDOn2jNuFG
推荐理由：发布信息列出了 Qwen3.8-Flash 的参数量、激活规模与定价，可作为判断高效多模态 MoE 路线的具体参照。
MarkTechPost精选AI 评分7676 阿里 Qwen 团队发布 Qwen3.8-Flash-Next，125B 多模态 MoE 预览 Qwen4 架构
阿里 Qwen 团队发布开源权重多模态 MoE 模型 Qwen3.8-Flash-Next，主模型 125B，每 token 仅激活 6B 参数，被定位为 Qwen4 架构的早期预览。
推荐理由：读者可借 6B 激活参数与训练成本对照，判断开源多模态模型在编码和智能体任务上的取舍。
歸藏的AI工具箱 · 微信公众号精选AI 评分7777 智谱 GLM-5.3 Flash 正式开源，匿名模型 Ox Alpha 身份揭晓
智谱揭晓匿名模型 Ox Alpha 的身份为 GLM-5.3 Flash，宣布 MIT 开源，API 已上架，Z Code 和智谱 Coding Plan 均可使用。
推荐理由：作者用三个有技术门槛的前端场景实测该模型与 DeepSeek V4 Flash 的同题表现，并给出与 Opus 4.8 的价格对比。
@Yuchenj_UW@Yuchenj_UW精选AI 评分7777 引用@Zai_org@Zai_orgIntroducing GLM-5.3-Flash - Leading capabilities at a highly competitive price - Natively multimodal with a 1M-token context window - A 320B-A18B model released under the MIT License - Previously previewed as Ox Alpha, running entirely on Chinese AI chips Blog: https://t.co/tzOmB7gdZP Available now across all official platforms: Weights: https://t.co/9LRMahY9Wa API: https://t.co/VcaQnzYmS9 Coding Plan: https://t.co/Nk8Y98HNhU ZCode: https://t.co/Peepqv4XSx Chat: https://t.co/WCqWT0qCQb AutoClaw: https://t.co/aGEG5HqTTb
推荐理由：原文对比了 GLM-5.3-Flash 与 GLM-5.2 的参数量和基准成绩，可据此了解高效小模型的进展。
@AYi_AInotes@AYi_AInotes精选AI 评分6868 
推荐理由：彭博社确认匿名模型 Ox Alpha 出自智谱并将在今晚开源权重，读者可了解它的真实来源与开放安排。
@kimmonismus@kimmonismus精选AI 评分6868 Zai 发布 GLM-5.3 Flash（Ox Alpha），320B MoE 每 token 仅激活 18B 参数，采用 MIT 开源许可，原生多模态并支持 1M 上下文。
引用@kimmonismus@kimmonismusThe upcoming Ox Alpha is GLM-5.3 Flash (as expected): 320b total parameters, 18b active. Outperforming GLM-5.2 at 1/10th of its price and approaching Opus 4.8 on coding and agentic benchmarks. Big things incoming! https://t.co/xZZmOD1Ghu https://t.co/H2fZ9idgVa
推荐理由：原文列出六项基准数据与 MIT 开源、1M 上下文等规格，便于读者判断这一稀疏 MoE 的效率定位。
IT Home精选AI 评分7676 智谱开源 GLM-5.3-Flash 原生多模态模型，限时折扣价为 GLM-5.3 的 1/20
智谱上线并开源 GLM-5.3-Flash（320B-A18B），这是 GLM-5 系列首个原生多模态模型，总参数量 320B、激活参数仅 18B。其在 Artificial Analysis Intelligence Index 取得 57 分，与 Claude Opus 4.8 持平，自研 Z.ai Code Bench 体感评估中编程表现也与之相当。
推荐理由：320B 总参数仅激活 18B 的架构设计搭配限时 1/20 定价，可供判断开源前沿模型的成本竞争区间。
Z.ai@Zai_org精选AI 评分7272
推荐理由：官方公告同时给出价格定位、开源权重和芯片适配信息，读者可以据此评估它在现有工作流中的替换可能。
@AYi_AInotes@AYi_AInotes精选AI 评分7676 Qwen 团队开源 Qwen3.8-Flash，总参数 125B 加 51B N-gram 嵌入，每 token 仅激活 6B，训练成本为 Qwen3.7-Plus 的 1/9。
引用@Alibaba_Qwen@Alibaba_Qwen⚡Meet Qwen3.8-Flash, a multimodal MoE and an early preview of the Qwen4 architecture, now open-weight! The production version Qwen3.8-Flash will be available soon via QwenCloud API at just $ 0.16/1M input tokens and $ 0.47/1M output tokens. 125B parameters + 51B N-gram embeddings, with just 6B activated per token. Unmatched cost-efficiency. What's new: 🥳 - Next architecture: GDN + QSA hybrid attention, Gated Residual, N-gram Embedding & Muon optimizer, serving as a precursor to the architecture used in Qwen4. - Dramatically lower training and inference costs: trained at just 1/9 the cost of Qwen3.7-Plus, while outperforming it across the board with especially strong gains in coding and office tasks. - Strong performance: scoring 58.7 on DeepSWE 1.1, 62.5 on SWE-bench Pro, 73.9 on CoWorkBench, 84.5 on AndroidWorld, and 95.7 on MathVision (with CI). - 262K native context, extensible to 1M with YaRN. We’re also releasing the weights for Qwen3.8-Flash-Next, giving the community an early look at the new architecture we’re exploring for Qwen4.🚀 We can't wait to see what you build with Qwen3.8-Flash!👀👇 - Blog: https://t.co/M5hYypFLgJ - Technical Report: https://t.co/IF0gObIkQO - Hugging Face: https://t.co/6ow8QVAABt - ModelScope: https://t.co/tDOn2jNuFG
推荐理由：Qwen3.8-Flash 以 125B 总参数、6B 激活和 1/9 训练成本给出开源 MoE 的效率样本，可对照其基准数据看架构取舍。
机器之心 · 微信公众号精选AI 评分7777 阿里发布 Qwen3.8-Flash，同步开源 Qwen3.8-Flash-Next
阿里发布 Qwen3.8-Flash，并在 Hugging Face 与 ModelScope 开放同一模型的 Qwen3.8-Flash-Next 权重，主模型 125B 参数、每 token 仅激活 6B，千问 AI 平台定价为每百万 token 输入 1 元、输出 3 元。
推荐理由：文章拆解了 Qwen3.8-Flash 在注意力、残差与嵌入上的四处架构改动，并把它放进每任务成本的行业对比框架里。
@kimmonismus@kimmonismus精选AI 评分8080 Qwen3.8-Flash-Next 发布，采用 125B MoE 参数加 51B N-gram embeddings，每 token 仅激活 6B 参数。
引用@kimmonismus@kimmonismusQwen 3.8 Flash-Next official released: A 6B-active open model just beat Claude Opus 4.6 Max across 8 of 9 comparable benchmarks! Qwen3.8-Flash-Next is a highly sparse MoE: • 125B model parameters • 51B additional n-gram embeddings • Only 6B parameters active per token It scores: • 62.5 SWE-bench Pro • 81.0 SWE-bench Multilingual • 73.9 CoworkBench • 55.7 JobBench • 73.5 Toolathlon • 81.3 IFBench • 91.7 GPQA Diamond • 91.9 LiveCodeBench It also outperforms Qwen3.8-27B and DeepSeek-V4-Flash across most of the table. Super cool release!!
推荐理由：原文给出四项架构改动与 1/9 训练成本的对比，读者可以了解高稀疏 MoE 如何压低单 token 计算量。
@kimmonismus@kimmonismus精选AI 评分7171 
推荐理由：6B 激活参数的稀疏 MoE 在多项基准上对标 Claude Opus 4.6 Max，可据此比较开源小激活模型的能力位置。
IT Home精选AI 评分8282 阿里通义发布并开源 Qwen3.8-Flash：125B 参数 MoE 模型，训练成本约为前代 1/9
阿里通义千问团队发布 Qwen3.8-Flash，并开源 Qwen3.8-Flash-Next 权重，官方称 Next 新架构将是全新一代 Qwen4 系列模型的雏形。
推荐理由：训练成本降到前代约九分之一，同时公开新架构与开源权重的具体设计，便于判断下一代 Qwen 的技术走向。
@Alibaba_Qwen@Alibaba_Qwen精选AI 评分7171 阿里通义千问发布多模态 MoE 模型 Qwen3.8-Flash 并开放权重，同时放出 Qwen3.8-Flash-Next 权重，作为其为 Qwen4 探索的新架构早期预览。
推荐理由：作为 Qwen4 架构的早期预览，文中给出了新架构组合与训练成本的具体对比数字。
Qwen Blog精选AI 评分6969 Qwen3.8-Flash-Next 开源，多模态 MoE 架构预览 Qwen4
千问团队开源 Qwen3.8-Flash-Next 权重，这是一个多模态 MoE 模型，也是 Qwen4 所用架构的早期预览。文中称其角色类似 Qwen3-Next 之于 Qwen3.5，当时的混合 Gated DeltaNet + Gated Attention 设计已用于 Qwen3.5 至 Qwen3.8 系列。
推荐理由：官方开源权重并定位为 Qwen4 架构预览，读者可据此追踪千问后续系列的架构走向。
Hugging Face Blog精选AI 评分7474 Sentence Transformers 教程：用 MultiVectorEncoder 训练与微调多向量嵌入模型
Hugging Face 发布 Sentence Transformers v6.0 教程，介绍第四种模型类型 MultiVectorEncoder 的完整训练方法，支持 ColBERT 风格的 late interaction 检索微调与从零训练。
推荐理由：作者用实测对比给出多向量模型微调的完整配方，包括起点选择、损失函数和索引压缩的量化取舍。
LMSYS Blog精选AI 评分7272 SGLang 为 Qwen3.8-Flash-Next 提供 Day-0 支持
SGLang 团队与 Qwen、NVIDIA、AMD 合作，为 Qwen 团队开源的多模态 MoE 模型 Qwen3.8-Flash-Next 提供 day-0 支持。
推荐理由：原文详解了 SGLang 对 Qwen3.8-Flash-Next 的 day-0 支持与架构实现细节，含量化检查点和实测吞吐数据，对部署选型有参考价值。
Hugging Face Blog精选AI 评分6363 IBM 发布 Granite 4.2 推理模型家族并详解构建过程
IBM 发布 Granite 4.2 密集 decoder-only 推理模型家族，含 3B、8B、30B 三个规格，基于 Granite-4.1 基座（约 15T tokens 预训练，上下文窗口扩至 512K），经 SFT 与多阶段 GRPO 强化学习训练，全部以 Apache 2.0 许可开源。
推荐理由：IBM 官方详解 Granite 4.2 训练全程，从五阶段预训练到多阶段 RL 课程，可复用的训练细节较完整。
@Alibaba_Qwen@Alibaba_Qwen精选AI 评分6767 引用@natolambert@natolambertOver the weekend I had Codex parse 500K arXiv AI/ML papers since ChatGPT to understand which open models are used for research. In 2024, ~30% of papers mentioned an American open model and only 10% a Chinese model. Today, ~40% of papers mention a Chinese (open) LLM, and only 25-30% an American one. Chinese models are the default for research. Chinese mentions are still growing while American open models are stagnating. When looking at this data it's important to remember that papers substantially lag model releases, as research takes a long time. Qwen's steady growth is reflective of this, but so is Llama's lasting power. Some more observations: 1. Qwen has been steadily growing, and today 1/3 of papers which mention any LLM mention qwen. OpenAI's closed models are the highest overall, at ~37%. 2. Llama peaked around April of 2025 at 30% of papers which mention any LLM (including ChatGPT etc). Llama 4 was released at about the same time, and Llama has been declining since. 3. Gemini and Claude are less common than the leading open models, mentioned in 10-15% of papers puts them behind all of Qwen, Llama, and DeepSeek. Open models should be and are the foundations of open research. The % of papers mentioning any LLM have been steadily climbing since 2023. | Year | January | April | July | October | | 2023 | 10.43% | 15.39% | 18.69% | 32.18% | | 2024 | 29.70% | 33.93% | 35.70% | 44.25% | | 2025 | 39.23% | 45.28% | 44.94% | 53.52% | | 2026 | 55.49% | 57.26% | 53.14% | TBD Now over 50% of AI papers, from 10% in 2023. Other notes: - Gemma and Mistral hover around 5-10%. - Our beloved fully-open Olmo models have been ~1% since the first release in Jan. 2024. - DeepSeek has a clear jump after R1 in Jan. 2025 - Data derived from the most popular ML arXiv categories: cs. AI, cs. CL, cs. CV, cs. LG, stat. ML Just like our downloads and derivative model data, this is updated daily on the Interconnects Open Model Dashboard.
推荐理由：引用数据呈现了近三年论文提及开源模型的份额变化，读者可据此观察中美开源模型在研究社区中的位置。