Hi NVIDIA Team,
I am a developer building autonomous agent swarms using OpenClaw and n8n on Ubuntu. My project involves multi-step reasoning and tool-calling using Llama 3.3 70B and Kimi k2.5.
The issue: Currently, my agents are hitting the 40 RPM limit almost immediately during the “thinking” and “searching” phases. This causes the Docker container tasks to fail with 429 errors.
Request: I am requesting a limit increase to 200 RPM to support these agentic loops and allow for stable testing of your MoE models.
Account Email: negraislam@gmail.com
Thanks for the support!