Request for NIM API Rate Limit Increase: 40 → 200 RPM for Multi-Agent AI Development Workflow

Hello NVIDIA Support Team,

I would like to request a rate limit increase for my NVIDIA NIM API account.

Account Details

  • Email: anddev888@gmail.com
  • Current limit: 40 RPM
  • Requested limit: 100–200 RPM (whatever is the next available tier for individual developers)

Use Case

I am using NVIDIA NIM API for personal development and non-production testing of multi-agent AI workflows. My setup integrates multiple LLM agents (coding agent, vision agent, planning agent) with local development tools for creative and software projects including ComfyUI pipelines, Blender automation, and GIMP plugin development.

A typical workflow involves:

  • Multi-agent orchestration with parallel subagent calls
  • Vision analysis via multimodal models (Gemma 4, Qwen)
  • Code generation, debugging, and refactoring across multiple files
  • Model comparison and evaluation across NIM-hosted models

Why 40 RPM is Insufficient

Even with client-side throttling and exponential backoff, a single agentic session triggers 20–50 API calls within minutes. The 40 RPM global limit across all models causes frequent 429 Too Many Requests errors that interrupt multi-step workflows mid-execution.

Commitment

  • Personal development only — no production or public-facing service
  • Low concurrency usage, will continue respecting Retry-After headers
  • Full compliance with NVIDIA’s fair-use policy

Thank you for considering my request.

Best regards, Yong