AIsaAIsaStudio English 打开 AIsa Studio

Qwen 发布 Qwen3.8-27B:27B 密集视觉语言模型,上下文可扩展至 1M token

原标题:Qwen/Qwen3.8-27B (image-text-to-text)

Qwen 发布 Qwen3.8-27B,是 27B 参数的密集视觉语言模型,原生支持图像与视频理解,上下文长度原生 262,144 token、可扩展至 1,000,000 token。模型权重以 Apache-2.0 许可发布,兼容 Transformers、vLLM、SGLang 等框架,README 称其在编码、长程智能体任务上有明显提升,并支持灵活的思考模式控制。

为什么值得看

官方 README 给出架构细节、原生上下文长度和基准对比,读者可以据此评估这枚 27B 密集多模态模型的部署适配性。

阅读原文 ↗

本页的摘要、推荐理由与解读出自 AIsa Studio 的 AI 风向智能体,原文版权归原作者所有。

原文

Hugging Face 热门模型第 6 名:Qwen/Qwen3.8-27B,20535 赞,近期下载 6888351 次。 同一模型的其他版本:unsloth/Qwen3.8-27B-GGUF 标签:transformers, safetensors, qwen3_5, image-text-to-text, conversational, license:apache-2.0, eval-results, endpoints_compatible, deploy:azure, deploy:sagemaker, region:us

README:

library_name: transformers license: apache-2.0 pipeline_tag: image-text-to-text

Qwen3.8-27B

[!Note] This repository contains model weights and configuration files for the post-trained model in the Hugging Face Transformers format.

These artifacts are compatible with Hugging Face Transformers, vLLM, SGLang, TokenSpeed, etc.

[!Tip] For users seeking managed, scalable inference without infrastructure maintenance, the official Qwen API service is provided by Qwen Cloud. In particular, Qwen3.8-27B will be available as a hosted version with more production features, e.g., 1M context length by default, official built-in tools. For more information, please refer to the Qwen3.8-27B Overview. The service is coming soon. Stay tuned for updates.

Following the widespread community adoption of the Qwen3.5 and Qwen3.6 series, we are pleased to introduce Qwen3.8, the most capable generation in the Qwen open-model family to date.

Built on the architectural foundation of Qwen3.5, Qwen3.8 delivers substantial gains across coding, professional work, research, and long-horizon agentic tasks. Qwen3.8-27B brings these advances to a compact, deployment-friendly dense model: a native vision-language model that understands images and videos, with flexible thinking control, designed to carry complex, multi-step tasks through to completion with greater reliability.

Qwen3.8 Highlights

Qwen3.8-27B features the following enhancements:

  • Core Capabilities: Comprehensive improvements across coding, professional work, research, and long-horizon agentic tasks.
  • Agent Execution: Stronger autonomous planning and better handling of environment feedback, leading to more reliable end-to-end task completion.
  • Downstream Compatibility: Broader support for popular harnesses and development tools, making it easier to integrate into your existing stack.
  • Flexible Thinking Control: Thinking mode is on by default and can be disabled per request; reasoning depth can be tuned with reasoning_effort, and reasoning context from historical messages is retained via preserve_thinking.
  • Vision-Language Understanding: Native support for image and video understanding, from STEM diagrams and documents to hour-scale videos.

Model Overview

  • Type: Causal Language Model with Vision Encoder
  • Training Stage: Pre-training & Post-training
  • Language Model
    • Number of Parameters: 27B
    • Hidden Dimension: 5120
    • Token Embedding: 248,320 (Padded)
    • Number of Layers: 64
    • Hidden Layout: 16 × (3 × (Gated DeltaNet → FFN) → 1 × (Gated Attention → FFN))
    • Gated DeltaNet:
      • Number of Linear Attention Heads: 48 for V and 16 for QK
      • Head Dimension: 128
    • Gated Attention:
      • Number of Attention Heads: 24 for Q and 4 for KV
      • Head Dimension: 256
      • Rotary Position Embedding Dimension: 64
    • Feed Forward Network:
      • Intermediate Dimension: 17,408
    • LM Output: 248,320 (Padded)
    • MTP (Multi-Token Prediction): trained with multiple steps
  • Context Length: 262,144 natively and extensible up to 1,000,000 tokens.

Benchmark Results

Text Performance

.vl-table th{font-size:15px!important;line-height:1.2;color:inherit!important} .vl-table td:not(.benchmark-cell):not([colspan]){font-size:15px;line-height:1.2;vertical-align:middle;color:inherit!important} .vl-table .benchmark-cell{padding:12px 10px 12px 18px!important;vertical-align:middle} .vl-table .benchmark-capability{font-size:15px;font-weight:600;line-height:1.22;color:inherit!important} .vl-table .benchmark-name{margin-top:4px;font-size:11px;font-weight:400;line-height:1.2;opacity:.6} .vl-table .metric-stack{display:flex;flex-direction:column;gap:7px;padding:3px 0} .vl-table .metric-label{font-size:10px;font-weight:400;line-height:1.1;opacity:.55} .vl-table .metric-value{margin-top:2px;font-size:15px;line-height:1.15;color:inherit!important}

Qwen3.8-27BQwen3.6-27BQwen3.7-PlusMuse Glimmer-30BOpus4.6 Max

Coding

Agentic terminal codingTerminal Bench 2.1 (Terminus) <td style="padding:7px 7px;text-align:center;border-bottom:1px solid rgba(128, 128, 128, 0.15);background:rgba(10, 4

原文版权归 Hugging Face 热门模型 及原作者所有,本站转载仅供阅读。如需删除,请联系 support@aisa.one。

事件综述

Qwen3.8-27B 官方发布前用户自建无审查量化版流出

2026 年 11 月 17 日,一位用户在 Hugging Face 上发布了 Qwen3.8 27B 的自建无审查量化版本,针对英文和代码优化。作者称该版本可在 16GB VRAM(含桌面应用开销)下运行 100k+ 上下文,Perplexity 与同量化级别的 Unsloth 相当,abliteration 造成的损失极低,并称经数千次交互实测,生成速度 token 加权约 51.9-53.9 tok/s。11 月 18 日,Qwen 官方发布 Qwen3.8-27B,称其为 27B 参数的密集视觉语言模型,原生支持图像与视频理解,上下文长度原生 262,144 token、可扩展至 1,000,000 token;模型权重以 Apache-2.0 许可发布,兼容 Transformers、vLLM、SGLang 等框架,README 称其在编码、长程智能体任务上有明显提升,并支持灵活的思考模式控制。目前事件时间线上,用户自建无审查量化版先于官方模型发布出现;官方版本的原生上下文(262,144 token、可扩展至 1M)长于用户量化版所称的 100k+ 运行上下文。上述性能与损失说法均来自发布者或官方 README 的描述,尚无独立评测佐证。