Multimodal
Qwen3.5 9B
Alibaba's dense Qwen3.5 9B vision-language model with Jetson-specific checkpoints for Orin and Thor
Parameters 9B
Inputs
Text Image Video
Outputs
Text
Context Length 256K
License Apache 2.0
Precision
NVFP4 W4A16
Serve the model
Pick your Jetson, then copy the command and paste it into a terminal on the Jetson. New to Jetson? Setup guide
Run on your Jetson
Engine
Speculative decoding
Weights
Model Details
View on HuggingFace
Qwen3.5 9B is a dense vision-language model in the Qwen3.5 family aimed at stronger reasoning, visual understanding, and agentic behavior on Jetson. This entry uses a W4A16 checkpoint on Jetson Orin and an NVFP4 checkpoint on Jetson Thor.
Inputs and Outputs
Input: Text and images
Output: Text
Intended Use Cases
- Visual reasoning: Stronger multimodal reasoning over image and text inputs
- Image understanding: Detailed captioning, scene description, and analysis
- Tool calling: Native Qwen tool-call parsing in vLLM
- Agents: Local assistants and workflow automation
Additional Resources
- Original Model - Base Qwen3.5 9B checkpoint
- W4A16 Checkpoint - Jetson Orin checkpoint
- NVFP4 Checkpoint - Jetson Thor checkpoint
Performance on Jetson
2K input · 128 output tokensWriting and summarization · up to 256 output tokens
Concurrent requests
vLLMllama.cpp
- 34.7 vLLM on AGX Thor: 34.7 tok/sAGX Thor128 GB
- 16.8 vLLM on AGX Orin: 16.8 tok/sAGX Orin64 GB
- 3.5 llama.cpp on Orin Nano: 3.5 tok/sOrin Nano8 GB
Output tokens/s
vLLM
- 159 vLLM on AGX Thor: 159 tok/sAGX Thor128 GB
- 55.3 vLLM on AGX Orin: 55.3 tok/sAGX Orin64 GB
Output tokens/s
vLLM
- 45.7 MTP vLLM with MTP on AGX Thor: 45.7 tok/sAGX Thor128 GB
- 30.1 MTP vLLM with MTP on AGX Orin: 30.1 tok/sAGX Orin64 GB
Output tokens/s
vLLM
- 291 MTP vLLM with MTP on AGX Thor: 291 tok/sAGX Thor128 GB
- 144 MTP vLLM with MTP on AGX Orin: 144 tok/sAGX Orin64 GB
Output tokens/s