Skip to content

Qwen 3.5

by Alibaba · Website

Alibaba's Qwen 3.5 is a multimodal model family spanning 0.8B to 397B parameters, supporting 201 languages with 256K context. Uses Gated Delta Networks and sparse MoE for efficient inference. All models support vision input and thinking/non-thinking modes. Released under Apache 2.0.

Variants (7)

Smallest:Qwen 3.5 0.8B (0.8B)
Largest:Qwen 3.5 122B (122B)

Qwen 3.5 0.8B

800M

Alibaba

Min 1.5 GB
text-generationreasoningmultilingual

Qwen 3.5 2B

2B

Alibaba

Min 3 GB
text-generationcode-generationreasoning

Qwen 3.5 4B

4B

Alibaba

Min 4.5 GB
text-generationcode-generationreasoning

Qwen 3.5 9B

9B

Alibaba

Min 7.5 GB
text-generationcode-generationreasoning

Qwen 3.5 27B

27B

Alibaba

Min 19 GB
text-generationcode-generationreasoning

Qwen 3.5 35B A3B

35B

Alibaba

Min 12 GB
text-generationcode-generationreasoning

Qwen 3.5 122B

122B

Alibaba

Min 85 GB
text-generationcode-generationreasoning