# NIM API rate limit increase request (40 → 200 RPM) — planning multi-model comparison setup

**URL:** <https://forums.developer.nvidia.com/t/nim-api-rate-limit-increase-request-40-200-rpm-planning-multi-model-comparison-setup/371282>\
**Category:** Access/Accounts\
**Tags:** nim, deepseek\
**Created:** [May 25, 2026, 1:23pm UTC](https://forums.developer.nvidia.com/t/nim-api-rate-limit-increase-request-40-200-rpm-planning-multi-model-comparison-setup/371282 "2026-05-25T13:23:42Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![sarkadi.csaba.zx](https://developer.download.nvidia.com/images/forums/profile-default-devtalk-84.png) [@sarkadi.csaba.zx](https://forums.developer.nvidia.com/u/sarkadi.csaba.zx)\
**Post date:** [May 25, 2026, 1:23pm UTC](https://forums.developer.nvidia.com/t/nim-api-rate-limit-increase-request-40-200-rpm-planning-multi-model-comparison-setup/371282/1 "2026-05-25T13:23:42Z")

</div>

Hi NVIDIA Developer Support,

I just got my API key set up on [build.nvidia.com](http://build.nvidia.com) and I’m planning a small

multi-model evaluation project — comparing how Kimi K2.6, GLM-5.1,

DeepSeek V4 Pro, and Step-3.5-Flash handle the same set of reasoning and

prompt-engineering test cases.

From reading the docs and looking at how similar projects are structured,

I can already see that the default 40 RPM will be the bottleneck the

moment I run a parallel batch across 4-6 models per prompt. Rather than

hitting 429s right out of the gate, I’d like to request an upfront

increase to 200 RPM.

Strictly personal / non-commercial use. Happy to provide more context

about the project if useful.

Thanks for considering!