To get started and build your AI app with NVIDIA NIM first you need to sign up with your either coporate or personel email.
| The benefits If you use a Corporate Email | The benefits If you use a Personal Email |
| Try NVIDIA NIMs with hosted endpoints | Try NVIDIA NIMs with hosted endpoints |
| 5000 credits for API Calls | 1000 credits for API Calls |
| Run NIMs on your infrastructure | |
| Enterprise Support for 90 days | |
| Security and vulnerability reports |
Build With Llama 3.1 Every Step of the Way Using NVIDIA AI
The Llama 3.1 collection of open models is now optimized with NVIDIA TensorRT-LLM for latency and throughput, and packaged as NVIDIA NIM™ inference microservices for the over 100 million GPUs being used worldwide—from data center to cloud to workstations.
Additionally, enterprises can use the NVIDIA AI Foundry, a platform and a service, which integrates Llama 3.1, to rapidly build custom Llama 3.1 “supermodels“’ with enterprise data and domain-specific use cases.
Building these supermodels requires high-quality domain data that can be generated using the Llama 3.1 405B and leaderboard-topping Nemotron-4 340B Reward models.
With NVIDIA NeMo™, developers can curate data, customize and evaluate foundation models, put guardrails in place, and generate high-accuracy responses based on up-to-date information using retrieval-augmented generation.
Enterprises can then pair the Llama 3.1 NIM with new NVIDIA NeMo Retriever to create state-of-the-art retrieval pipelines for generative AI copilots, AI assistants, and digital human avatars.
To supercharge enterprise deployments for production AI, Llama 3.1 NIMs are now available for download from ai.nvidia.com.
