//
Run Alibaba's powerful open-source LLM
Qwen 2.5 is Alibaba's latest open-source LLM family, offering models from 0.5B to 72B parameters. Known for excellent multilingual support and strong coding abilities.
Jan has Qwen models in its Model Hub.
# Download: https://jan.ai/downloadSearch 'Qwen' to see all available versions.
Qwen 2.5 sizes:
• 0.5B, 1.5B, 3B - Ultra lightweight
• 7B, 14B - Sweet spot for most users
• 32B, 72B - Maximum capability💡 Qwen excels at Chinese, coding, and math.
Choose based on your GPU VRAM.
8GB VRAM: Qwen 2.5 7B Q4
12GB VRAM: Qwen 2.5 14B Q4
16GB VRAM: Qwen 2.5 32B Q4
24GB VRAM: Qwen 2.5 72B Q4❓ Chinese characters display incorrectly
✅ Ensure your terminal/app supports Unicode. Jan handles this automatically.
❓ Qwen 72B is too slow
✅ 72B needs RTX 4090 24GB for reasonable speeds. Try 32B on 16GB cards.