Qwen3
Demo Environment
Development Board : Jetson Orin Series Motherboard
Due to performance limitations, the Jetson Orin Nano 4GB must run the reduced-parameter version
Qwen 3 is the latest generation of large-scale language models in the Qwen series, providing a complete suite of dense and mixture-of-experts (MoE) models.
1. Model Size
| Model | Size |
|---|---|
| qwen3:0.6b | 523MB |
| qwen3:1.7b | 1.4GB |
| qwen3:4b | 2.6GB |
| qwen3:8b | 5.2GB |
2. Performance


3. Using Qwen3
3.1. Running Qwen3
Use the run command to start running the model. If you haven't downloaded the model, the model from the Ollama model library will be automatically downloaded:
ollama run qwen3:8b

3.2. Starting a Conversation
Please tell me how many hours there are in a day.
Response time depends on your hardware configuration, so please be patient!

3.3. Ending the Conversation
Use the Ctrl+d shortcut or /bye to end the conversation!
References
Ollama
Official Website: https://ollama.com/
GitHub: https://github.com/ollama/ollama
Qwen3
GitHub: https://github.com/QwenLM/Qwen3
Ollama Model: https://ollama.com/library/qwen3