Community · Oct 25, 2026, 06:10 UTC

Developer previews a 4B coding model fine-tuned from Qwen3.5-4B for 3D scene generation

An independent fine-tune of Qwen3.5-4B aimed at coding and 3D scene generation, with Q4_K_M and Q8_0 GGUFs on Hugging Face. Training is still running, and the training dataset is promised later.

• LabAlibaba
• ModelQwen3.5-4B
• Topicopen-weights
• Topiccoding
• Topicproduct