NVIDIA NIM

Inference layers 10 models data_array 262K max context Free tier

info Notes

Free with NVIDIA Developer Program membership. 100+ models. Rate-limited (no daily token cap).

key How to get API key

General steps — follow the official NVIDIA NIM page for exact details.

  1. Open the NVIDIA NIM API key page ↗
  2. Create a new account, or sign in if you already have one.
  3. Generate an API key from the dashboard or settings.

layers Available models 10

search
Model Capabilities Context Output Rate limit
`deepseek-ai/deepseek-r1` `deepseek-ai/deepseek-r1`
Text (reasoning)
128K ~40 RPM
`nvidia/nemotron-3-super-120b-a12b` `nvidia/nemotron-3-super-120b-a12b`
description
262K ~40 RPM
`nvidia/nemotron-3-nano-30b-a3b` `nvidia/nemotron-3-nano-30b-a3b`
description
128K ~40 RPM
`nvidia/llama-3.1-nemotron-ultra-253b-v1` `nvidia/llama-3.1-nemotron-ultra-253b-v1`
description
128K ~40 RPM
`meta/llama-3.1-405b-instruct` `meta/llama-3.1-405b-instruct`
description
128K ~40 RPM
`qwen/qwen2.5-72b-instruct` `qwen/qwen2.5-72b-instruct`
description
128K ~40 RPM
`google/gemma-4-31b` `google/gemma-4-31b`
description
128K ~40 RPM
`mistralai/mistral-large-2-instruct` `mistralai/mistral-large-2-instruct`
description
128K ~40 RPM
`minimax/minimax-m2.7` `minimax/minimax-m2.7`
description
128K ~40 RPM
+ 90 more models +-90-more-models
Text, Image, Video, Speech, Embeddings
Varies ~40 RPM

help Frequently Asked Questions

Yes! NVIDIA NIM provides free API access or free developer credits. Refer to the model list table above for context window and rate limit details.

Most free tier providers in our directory (including NVIDIA NIM) do not require a credit card for developer key activation. Check their official developer portal for exact signup requirements.

If your application exceeds the RPM (Requests Per Minute) or RPD (Requests Per Day) quota, the API will respond with an HTTP 429 status code until your quota window resets.