中文简介
Qwen3 30B-A3B 是混合专家（MoE）模型，总参数约 30.5B、每次推理激活约 3.3B。它旨在用较低激活参数获得更强容量，适合研究 MoE 性能、服务器推理和质量效率平衡的团队。
上游模型卡 / 数据集卡
Qwen3 30B-A3B 是混合专家模型。官方模型卡给出的规模约为 30.5B 总参数、3.3B 激活参数；两者含义不同，激活参数不能直接等同于存储或内存需求。
本版本原生上下文长度为 32,768 tokens，可按官方 YaRN 说明扩展至最高 131,072 tokens，并支持思考与非思考模式。
部署前应确认推理框架的 MoE 兼容性，评估完整权重、设备间通信、上下文和并发对资源的影响。许可证与配置以上游固定版本为准。
Qwen3-30B-A3B Qwen3 Highlights Qwen3 is the latest generation of large language models in Qwen series, offering a comprehensive suite of dense and mixture-of-experts (MoE) models. Built upon extensive training, Qwen3 delivers groundbreaking advancements in reasoning, instruction-following, agent capabilities, and multilingual support, with the following key features: **Uniquely support of seamless switching between thinking mode** (for complex logical reasoning, math, and coding) and **non-thinking mode** (for efficient, general-purpose dialogue) **within single model**, ensuring optimal performance across various scenarios. **Significantly enhancement in its reasoning capabilities**, surpassing previous QwQ (in thinking mode) and Qwen2.5 instruct models (in non-thinking mode) on mathematics, code generation, and commonsense logical reasoning. **Superior human preference alignment**, excelling in creative writing, role-playing, multi-turn dialogues, and instruction following, to deliver a more natural, engaging, and immersive conversational experience. **Expertise in agent capabilities**, enabling precise integration with external tools in both thinking and unthinking modes and achieving leading performance among open-source models in complex agent-based tasks. **Support of 100+ languages and dialects** with strong capabilities for **multilingual instruction following** and **translation**. Model Overview Qwen3-30B-A3B** has the following features: Type: Causal Language Models Training Stage: Pretraining & Post-training Number of Parameters: 30.5B in total and 3.3B activated Number of Paramaters (Non-Embedding): 29.9B Number of Layers: 48 Number of Attention Heads (GQA): 32 for Q and 4 for KV Number of Experts: 128 Number of Activated Experts: 8 Context Length: 32,768 natively and 131,072 tokens with YaRN. For more details, including benchmark evaluation, hardware requirements, and inference performance, please refer to our blog, GitHub, and Documentation. Quickstar
适用场景
适合较复杂的推理、代码、知识工作和 MoE 部署评测。是否优于同级稠密模型，应在目标框架、硬件和并发条件下用真实任务比较。
模型参数
混合专家模型，总参数约 30.5B、激活参数约 3.3B；原生上下文长度为 32,768 tokens，可按 YaRN 说明扩展至最高 131,072 tokens。支持思考与非思考模式。
文件说明
固定版本的官方 API 元数据显示 26 个文件条目，总存储量约 56.88 GB。MoE 的激活参数不能用于推断仓库体积，交付和部署都需按完整文件集合评估。
上游文件元数据
.gitattributes1.53 KBconfig.json963 Bgeneration_config.json239 BLICENSE11.08 KBmerges.txt1.59 MBmodel-00001-of-00016.safetensors3.72 GBmodel-00002-of-00016.safetensors3.73 GBmodel-00003-of-00016.safetensors3.72 GBmodel-00004-of-00016.safetensors3.73 GBmodel-00005-of-00016.safetensors3.73 GBmodel-00006-of-00016.safetensors3.73 GBmodel-00007-of-00016.safetensors3.73 GB
硬件建议
需要能支持完整 MoE 权重与目标精度的环境，并关注框架兼容性、设备间带宽、KV Cache 和并发。建议先验证推理引擎是否明确支持该架构，再做容量与性能测试。
注意事项
激活参数少不代表只需存储或加载激活部分，完整专家权重仍会影响磁盘、内存、启动和传输。框架对 MoE 的支持、路由效率和跨设备通信也会影响结果。
本页面为橙子AI科技基于固定版本上游卡片整理的中文信息与服务说明，不代表资源作者或平台官方页面。实际许可、访问和使用条件以上游原文为准。