deepseek 满血版部署

how does deepseek-r1 compare to other reasoning models