NIM API Rate Limit Increase: 200 RPM (or max research tier) for Agentic Workflows

hi NVIDIA Team,

I am currently developing a multi-agent orchestration framework at Tbilisi State University and have reached the limits of the default 40 RPM tier.

My research focuses on Agentic Workflows where multiple LLM agents (primarily using DeepSeek V4 and Llama 3.3) perform parallel reasoning and cross-verification. Because these agents operate concurrently, the current 40 RPM limit results in immediate 429 “Too Many Requests” errors, which disrupts the logic of the agentic chains.

I am requesting an increase to 200 RPM, or the highest available tier for independent academic research, to support this high-concurrency testing and to begin integrating FLUX.1 for visual asset generation within the workflow.

Account Details:

  • Email: kavadzes@gmail.com

  • API Key ID (last 4): ZssM

  • Current Limit: 40 RPM

  • Requested Limit: 200+ RPM

Thank you for your support of the developer and research community.

Best regards,
[User]

This forum is not for requesting rate limit increases. Also 40 RPM is a very generous limit, it’s way more than enough for development…

Hi NVIDIA Team, I realized that rate limit increases aren’t granted here. If this request is not possible, please delete this topic for me. Thank you for your help!