Responses very slow on nvidia-nim

[arsh0429]

{Free tier}
Can we only use the free-endpoint models mentioned in nvidia-nim because I cannot use glm 5.1 or deepseek-v4-pro models? Also I am currently using minimaxai/minimax-m2.7 model and it takes around 2 mins to answer even a simple query as "hi", I am using /effort to be medium- Kindly help and suggest me tips to make it faster

[soulfruit2010-cell]

yes, same issues very not stable with nvidia nim as default model