DeepSeek R1 0528 - 8B vs 671B Logic Reasoning Test

This video demonstrates a comparative analysis of the distilled DeepSeek R1 0528 Qwen3 8B model against its full 671B counterpart, specifically testing their reasoning capabilities. Watch as the presenter conducts an "elevator test" logic puzzle to highlight performance differences between these AI models. The demonstration begins with an introduction to both models, followed by a detailed logic test of the 671B model with multiple runs to assess consistency. Later, the same test is applied to the 8B distilled model, with evaluation runs to determine its reasoning limitations. The video provides timestamps for each segment, including the initial test, multiple runs of both models, and final optimization attempts for the 8B model, offering valuable insights into how model size affects reasoning performance.