@TomNVIDIA there’s nothing we can do? With the strong of our community? NVIDIA NIM API is completely useless right now! I don’t think this is because NVIDIA can’t handle this, but because of this amount of access without any limitations. Hope NVIDIA NIM don’t die soon.
that is the second admin/moderator pinged but seem nvidia will not answer …
Omg, so it’s just a matter of time to this plataform actually die.
same thing ban openclaw
we were fighting to ban openclaw for over 1 months now, but it seem nvidia not care about this
I suppose some AI agents are also token burners, when not configured to be optimized. It can easily burn through billions of tokens if you’re not careful. When you’re using paid services, you usually set some limits to not drain your wallet by accident, which users probably aren’t doing here.
Some people also set extra/fake accounts to rotate, as 40 rpm is usually not good enough in these agents.
Some of the more sought models like: glm-5.1, Kimi k-2.6, deepseek-v4-pro, deepseek-v4-flash, minimax-m2.7 are all unstable (constant Time Out or Internal Server Error) or slower than normal right now.
I’m actually surprised glm-5.1 is unstable, as glm-4.7 used to be very fast. Same for Kimi k-2.6 predecessors. But again, it’s been months since I’ve last used the service.
Nemotron is good for now, at least.
Yeah i agree. I feel like Open Claw users are overusing Nvidia Nim. They’re grabbing all the free tier capacity, which negatively impacts fair use, causing slow responses and even errors.