ModelDeepSeek-R1-Distill-Qwen-7B

DeepSeek-R1-Distill-Qwen-7B

A 7-billion-parameter language model distilled from DeepSeek's R1 reasoning model, built on the Qwen architecture. It is designed with a focus on mathematical and reasoning capabilities and represents the lightweight reasoning variant within DeepSeek's open-source model series.

Timeline (last 90 days)

Sep 18

Using Reflective Recovery, DeepSeek-R1-Distill-Qwen-7B's accuracy on AIME 2025 improved from 30.0% to 37.5%.

Unverified50%

All Facts (1)

Unverified

Using Reflective Recovery, DeepSeek-R1-Distill-Qwen-7B's accuracy on AIME 2025 improved from 30.0% to 37.5%.

50%

Source Articles

Reflective Recovery: Teaching LLMs to Self-Correct by Learning from MistakesSep 18