DeepSeek-R1-Distill-Qwen is a distilled reasoning-focused language model derived from DeepSeek-R1 and adapted to a Qwen backbone, retaining strong logical reasoning ability while reducing model size and inference cost.

Original paper: DeepSeek-R1

DeepSeek-R1-Distill-Qwen-7B

This model implements the DeepSeek-R1-Distill-Qwen variant, optimized to deliver efficient reasoning, coding, and analytical capabilities with practical latency and compute requirements. It is well suited for applications such as coding assistants, mathematical reasoning, structured analysis, and enterprise AI copilots where strong reasoning is needed in deployable form factors.

Model Configuration:

• Reference implementation: DeepSeek-R1 project repository
• Original Weight: DeepSeek-R1-Distill-Qwen-7B
• Support Cooper version:
  • Cooper SDK: [2.5.3]
  • Cooper Foundry: [2.2]

Model	Device	Model Link
DeepSeek-R1-Distill-Qwen	N1-655	Model_Link