From the model card
Fields below are copied from the tags and counters on the HuggingFace repository Qwen/Qwen3.5-27B at our last fetch. They are set by the uploader, not verified by us; rows with no tag are omitted. How this page is made.
- Publisher (HF namespace)
- Qwen
- Pipeline tag
- image-text-to-text
- Library
- Transformers
- Weight formats
- safetensors
- License tag
apache-2.0— read the license file in the repo before relying on it- Downloads (HF counter at last fetch)
- 2,369,643
- Likes (HF counter at last fetch)
- 1,040
- Model card
- https://huggingface.co/Qwen/Qwen3.5-27B
Use cases
- Multimodal document and chart analysis
- Complex multi-step reasoning with image context
- High-quality bilingual Chinese-English content generation
- Fine-tuning base for domain-specific vision-language tasks
Pros
- Apache-2.0 licensed
- Supports both image and text input natively
- Larger capacity than 14B class models for knowledge-intensive tasks
- Strong multilingual performance from Alibaba's training data
Cons
- 27B in bfloat16 requires ~54GB VRAM — needs multi-GPU or quantization
- Instruction following can be verbose without explicit length constraints
- Benchmark results show quality gaps vs GPT-4o on spatial reasoning
- Limited third-party quantization support compared to Llama/Mistral
Tags
transformerssafetensorsqwen3_5image-text-to-textconversationallicense:apache-2.0eval-resultsendpoints_compatibleregion:usdeploy:sagemakerdeploy:azure