Instantly Deploy Generative AI With NVIDIA NIM

Explore the latest community-built AI models with an API optimized and accelerated by NVIDIA, then deploy anywhere with NVIDIA NIM™ inference microservices.

Experience Leading Open Models Now


Accelerated AI is Just an API Call Away

Get up and running quickly with familiar APIs.

Seamless Compatibility With Popular Libraries

Use NVIDIA APIs from your existing tools and applications with as little as three lines of code.

Build With the Tools You Love

Work with your favorite LLM programming frameworks, including LangChain and LlamaIndex, and easily deploy your applications.

Run Anywhere

Accelerate Your AI Deployment With NVIDIA NIM

Part of NVIDIA AI Enterprise, NVIDIA NIM is a set of easy-to-use inference microservices for accelerating the deployment of foundation models on any cloud or data center and helping to keep your data secure.

Accelerate Your AI Deployment With NVIDIA NIM

Deploy NIM

Deploy NIM for your model with a single command. You can also easily run NIM with fine tuned-models.

Run Inference

Get NIM up and running with the optimal runtime engine based on your NVIDIA-accelerated infrastructure.


Developers can integrate self-hosted NIM endpoints in just a few lines of code.

docker run nvcr.io/nim/publisher_name/model_name
curl -X 'POST' \ '' \ -H 'accept: application/json' \ -H 'Content-Type: application/json' \ -d '{ "model" : "model_name", "prompt" : "Once upon a time", "max_tokens" : 64 }'
import openai client = openai.OpenAI( base_url = "YOUR_LOCAL_ENDPOINT_URL", api_key="YOUR_LOCAL_API_KEY" ) chat_completion = client.chat.completions.create( model="model_name", messages=[{"role" : "user" , "content" : "Write me a love song" }], temperature=0.7 )

Launch Locally or Scale With Kubernetes

Seamlessly deploy containerized AI microservices on any NVIDIA accelerated infrastructure, from a single device to data center scale.

Deploy Securely With Confidence

Rely on production-grade runtimes, including ongoing security updates, and run your business applications with stable APIs backed by enterprise-grade support.

Lower Costs and Your Carbon Footprint

Lower the operational cost of running models in production with AI runtimes that are continuously optimized for low latency and high throughput on NVIDIA-accelerated infrastructure.


NVIDIA NIM provides optimized throughput and latency out of the box to maximize token generation, support concurrent users at peak times, and improve responsiveness.

Configuration: Llama 3-70b-instruct, input token length: 7,000, output token length: 1,000. Concurrent client requests: 100, on 4 x H100 SXM NVLink. NIM OFF: FP16, TTFT: ~120s, ITL: ~180ms; NIM ON: FP8, TTFT: ~4.5s, ITL: ~70ms.

See NVIDIA NIMs in Action

Get Started

Start Prototyping for Free

Get started with easy-to-use, NVIDIA-managed serverless APIs.

  • Access fully accelerated AI infrastructure.
  • Ensure your data isn't used for model training.
  • Get started for free with 1,000 inference credits.

Download and Deploy

Run NVIDIA NIM to scale optimized AI models in the cloud or data center of your choice.

  • Ensure that data never leaves your secure enclave.
  • Seamlessly transition from cloud endpoints to self-hosted APIs without code changes.
  • Use an NVIDIA AI Enterprise license for production, or get started for free with the NVIDIA Developer Program.

Use Cases

Ignite Your Innovation

See how NVIDIA APIs support industry use cases and jump-start your AI development with curated examples.

Digital Humans

Bring game characters to life or create interactive virtual avatars to enhance customer service, empowering your application to connect more deeply with users.

Content Generation

Generate highly relevant, bespoke, and accurate content, grounded in the domain expertise and proprietary IP of your enterprise.

Biomolecular Generation

Biomolecular generative models and the computational power of GPUs efficiently explore the chemical space, rapidly generating diverse sets of small molecules tailored to specific drug targets or properties.


Take Your Enterprise AI Farther, Faster

Join leading partners to develop your AI applications with models, toolkits, vector databases, frameworks, and infrastructure from our ecosystem.


Unlock, Upskill, and Upscale

NVIDIA LaunchPad

Unlock AI With a Hands-On Lab

Experience end-to-end AI solutions through guided hands-on labs for development frameworks, retrieval-augmented generation (RAG)-based chatbots, route optimizations, and more.

NVIDIA Developer Program

Accelerate Your AI Applications

Get free access to NIM for application development, research, and testing plus technical learning resources through the NVIDIA Developer Program.

AI Workbench

Start Small. Scale Big

NVIDIA AI Workbench gives developers the flexibility to run API-enabled models on local or remote GPU-powered containers, allowing for interactive project workflows from experimentation to prototyping to proof of concept.


Explore NVIDIA NIM in the News

Check out the latest NVIDIA press releases to see how NIM and generative AI are impacting industries, partners, customers, and more.


Explore technical documentation to start prototyping and building your enterprise AI applications with NVIDIA APIs, or scale on your own infrastructure with NVIDIA NIM.