趋近智
Qwen3.5-2B
参数
2B
上下文长度
262.144K
模态
Multimodal
架构
Dense
许可证
Apache 2.0
发布日期
24 Feb 2026
训练数据截止日期
-
技术规格
注意力结构
Grouped-Query Attention
隐藏维度大小
2048
层数
24
注意力头
8
键值头
2
激活函数
SwigLU
归一化
RMS Normalization
位置嵌入
ROPE
Qwen3.5-2B
Qwen3.5-2B is Alibaba Cloud's small-scale multimodal foundation model with 2B parameters, released February 2026. It uses a hybrid architecture combining Gated Delta Networks and Gated Attention in a 6×(3×DeltaNet→FFN→1×Attention→FFN) pattern. In thinking mode, it achieves MMLU-Pro (74.0%), GPQA Diamond (65.8%), and GPQA (51.6%). Features unified vision-language capabilities, 262k native context, multi-token prediction training, and supports both thinking and non-thinking modes for prototyping, fine-tuning, and research purposes across 201 languages.
关于 Qwen 3.5
Qwen 3.5 is Alibaba Cloud's latest-generation foundation model family, released February 2026. It represents a significant leap forward, integrating breakthroughs in multimodal learning (unified vision-language foundation), efficient hybrid architecture (Gated Delta Networks with sparse Mixture-of-Experts), scalable reinforcement learning across million-agent environments, and global linguistic coverage spanning 201 languages. Available under Apache 2.0 license with open weights.
其他 Qwen 3.5 模型
评估基准
没有可用的 Qwen3.5-2B 评估基准。