Nvidia NIM API Pricing (2026)
NVIDIA NIM (NVIDIA Inference Microservices) provides optimized AI model deployment through NVIDIA's infrastructure. Features Nemotron foundation models and optimized versions of open-source models, leveraging NVIDIA's TensorRT-LLM for maximum inference performance on NVIDIA GPUs. Founded in 2024 and headquartered in Santa Clara, CA, Nvidia NIM offers 6 active models through their API. All pricing data below is sourced directly from the Nvidia NIM pricing page and is updated regularly.
Prices verified Jun 26, 2026
Nvidia NIM Model Pricing — All Models
All prices are in USD per 1 million tokens ($/M tokens). Input tokens are the text you send to the model; output tokens are the generated response.
| Model | Tier | Input | Output | Cached Input | Batch Input | Batch Output |
|---|---|---|---|---|---|---|
Nemotron Super 49B NVIDIA Nemotron Super 49B is a high-efficiency foundation model from NVIDIA's Nemotron family, optimized via TensorRT-LLM for fast inference on NVIDIA GPUs. Delivers strong general-purpose performance at budget-tier pricing, suitable for production workloads requiring a balance of capability and cost. | Budget | $0.10/M | $0.40/M | — | — | — |
Nemotron 70B Instruct NVIDIA Nemotron 70B Instruct is a 70B parameter instruction-tuned model from NVIDIA's Nemotron family, built on Llama 3.1 architecture and fine-tuned for helpfulness and alignment. Optimized with TensorRT-LLM for efficient inference on NVIDIA GPUs with strong reasoning and instruction-following capabilities. | Mid | $1.20/M | $1.20/M | — | — | — |
Nemotron Nano 12B VL NVIDIA Nemotron Nano 12B VL is a compact 12B vision-language model from the Nemotron Nano family, capable of processing both text and images. Optimized with TensorRT-LLM for fast multimodal inference on NVIDIA GPUs, making it ideal for budget-conscious applications requiring visual understanding. | Budget | $0.20/M | $0.60/M | — | — | — |
Nemotron Nano 9B NVIDIA Nemotron Nano 9B is an ultra-compact 9B parameter model from the Nemotron Nano family, offering the lowest cost entry point in NVIDIA's NIM portfolio. TensorRT-LLM optimized for maximum throughput on NVIDIA GPUs, ideal for high-volume, latency-sensitive applications at minimal cost. | Budget | $0.04/M | $0.16/M | — | — | — |
Nemotron 3 Super Nemotron 3 Super by Nvidia NIM. | Budget | $0.10/M | $0.50/M | — | — | — |
NVIDIA: Nemotron 3 Ultra NVIDIA: Nemotron 3 Ultra by Nvidia NIM. | Budget | $0.50/M | $2.20/M | — | — | — |
Prices last verified: 2026-06-26. Always confirm on the official pricing page before production use.
Nvidia NIM Model Capabilities
Key technical capabilities across all Nvidia NIM models. Use this table to identify which models support the features your application requires.
| Model | Context | Max Output | Vision | Functions | JSON Mode | Caching | Batch API | Reasoning |
|---|---|---|---|---|---|---|---|---|
| Nemotron Super 49B | 128,000 | 8,192 | — | ✓ | ✓ | — | — | — |
| Nemotron 70B Instruct | 128,000 | 8,192 | — | ✓ | ✓ | — | — | — |
| Nemotron Nano 12B VL | 128,000 | 8,192 | ✓ | ✓ | ✓ | — | — | — |
| Nemotron Nano 9B | 128,000 | 8,192 | — | ✓ | ✓ | — | — | — |
| Nemotron 3 Super | 262,144 | 4,096 | — | ✓ | ✓ | — | — | — |
| NVIDIA: Nemotron 3 Ultra | 1,000,000 | 16,384 | — | ✓ | ✓ | — | — | ✓ |
When to Choose Nvidia NIM
Cost-sensitive workloads
Use Nemotron Nano 9B — Nvidia NIM's most affordable model at $0.04/M input / $0.16/M output. Estimated cost at 100K requests/month: $12.00.
High-volume pipelines
Nvidia NIM models are well-suited for high-throughput workloads. Consider batch API (where supported) for an additional 50% cost reduction on asynchronous pipelines. Check the pricing table above for batch pricing availability.
Compare Nvidia NIM Models vs Competitors
See detailed cost comparisons between Nvidia NIM models and models from other providers, including side-by-side pricing tables and monthly cost projections.
Use Cases for Nvidia NIM Models
Browse use case guides where Nvidia NIM models are recommended, including cost-effective model rankings and monthly cost estimates for each workload.
Frequently Asked Questions: Nvidia NIM API Pricing
What is the cheapest Nvidia NIM model?
The cheapest Nvidia NIM model by input token price is Nemotron Nano 9B, priced at $0.04/M for input tokens and $0.16/M for output tokens. It is best suited for high-volume, cost-sensitive workloads where speed and price matter most.
What is the most capable Nvidia NIM model?
Nemotron Super 49B is Nvidia NIM's most capable model, classified as a budget-tier model. It supports a context window of 128,000 tokens and is priced at $0.10/M input / $0.40/M output per million tokens.
How does Nvidia NIM API pricing work?
Nvidia NIM charges per token consumed — separately for input tokens (the text you send) and output tokens (the text the model generates). Prices are listed in USD per 1 million tokens ($/M). Nvidia NIM was founded in 2024 and is headquartered in Santa Clara, CA. All pricing shown here is sourced from the official Nvidia NIM pricing page.
Does Nvidia NIM offer batch API pricing?
Nvidia NIM's current model lineup does not include batch API pricing at this time. For cost reduction on high-volume workloads, consider prompt caching (where supported) or running multiple concurrent requests with standard pricing.
Which Nvidia NIM models support vision / image inputs?
1 Nvidia NIM model support vision (image input): Nemotron Nano 12B VL. These models can analyze images, charts, screenshots, and documents alongside text prompts.
Nvidia NIM pricing data last verified: 2026-06-26. Prices may change — verify on the official Nvidia NIM pricing page before production use.