Free tier/Report for October 2026

NVIDIA NIM

Free API tier limits · LLM + multimodal inference

Free NIM API endpoints for prototyping via the NVIDIA Developer Program; production use needs an AI Enterprise license.

Card unverifiedVerified Oct 21, 20262 official sourcesLLM + multimodal inference
—requests / minute · unverified
—requests / day · not published
—tokens / minute · not published
Unknowncard requirement unverified

Free allowance

Free for prototyping, research, development and testing only.

Per-model limits shown in your build.nvidia.com account; vary by model and current traffic.

Credits. Separate 90-day NVIDIA AI Enterprise trial for production-grade self-hosting.

Models on the free offer

API catalog models on build.nvidia.com

Worth knowing

  • build.nvidia.com's help FAQ (as indexed by search engines) states 'up to 40 requests per minute (RPM) for most models'; the page is client-rendered and we could not read it directly, so the RPM field stays blank until verified.
  • Production use requires an NVIDIA AI Enterprise license (from $4,500/GPU/year per FAQ).
Left blank on purpose: requests per minute, whether a card is required. No official public source as of Oct 21, 2026.

Questions

Does NVIDIA NIM have a free API tier?
Free tier. Free NIM API endpoints for prototyping via the NVIDIA Developer Program; production use needs an AI Enterprise license.
Does NVIDIA NIM require a credit card?
We could not verify this from an official page, so we leave it blank rather than guess.
What are NVIDIA NIM's free rate limits?
Per-model limits shown in your build.nvidia.com account; vary by model and current traffic.
When was this NVIDIA NIM data last verified?
Oct 21, 2026, against: NIM FAQ; build.nvidia.com help.