NVIDIA API pricing & API key
Every NVIDIA model on one page: 3 models, from $0 to $0.22 per 1M input tokens. Get one NVIDIA API key from YepAPI and pay per request — no subscription, no NVIDIA account or waitlist.
Takes 30 seconds.
NVIDIA API pricing (3 models)
YepAPI prices per 1M tokens (text) or per output (media), cheapest first. Updated October 23, 2026.
| Model | Input / 1M | Output / 1M | Context |
|---|---|---|---|
| Nemotron 3 Super | $0 | $0 | 262K |
| Nemotron 3.5 Lightning | $0.10 | $0.30 | 262K |
| Nemotron 3.5 Super VL | $0.22 | $0.74 | 1.0M |
All NVIDIA models
Nemotron 3.5 Lightning
Text GenerationAccess Nemotron 3.5 Lightning through one API key. NVIDIA Open High-Throughput MoE.
Nemotron 3 Super
Text GenerationAccess Nemotron 3 Super through one API key. NVIDIA's free model.
Nemotron 3.5 Super VL
Text GenerationAccess Nemotron 3.5 Super VL through one API key. NVIDIA Open Multimodal MoE.
NVIDIA API through YepAPI
YepAPI provides access to NVIDIA models through a unified API endpoint. Instead of managing a separate NVIDIA API key and billing account, use your single YepAPI key to call any NVIDIA model alongside 282 other AI models from OpenAI, Anthropic, Google, Meta, and more.
All NVIDIA models use the same OpenAI-compatible request format. Switch between models by changing the model parameter — no SDK changes, no code rewrites. Pay per token with no monthly minimums.
How to get a NVIDIA API key
- Sign up at yepapi.com with Google or email — about 30 seconds.
- Open the dashboard and create an API key (it starts with yep_sk_).
- Call any NVIDIA model with that key — the OpenAI-compatible /v1/ai/chat/completions endpoint for text models.
- Top up any amount; usage is billed per request.
NVIDIA API FAQ
- How much does the NVIDIA API cost?
- NVIDIA models cost from $0 to $0.22 per 1M input tokens through YepAPI. The cheapest is Nemotron 3 Super ($0 in / $0 out per 1M). You pay per request with no monthly fee.
- How do I get a NVIDIA API key?
- Sign up at yepapi.com (about 30 seconds), open the dashboard and create an API key. The one key works for all 3 NVIDIA models and 282 models from other providers — no separate NVIDIA account.
- What is the cheapest NVIDIA model?
- Nemotron 3 Super — $0 in / $0 out per 1M.
- Which NVIDIA model has the largest context window?
- Nemotron 3.5 Super VL, with 1.0M tokens of context.
- Can I use the NVIDIA API with the OpenAI SDK?
- Yes. Set the base URL to https://api.yepapi.com/v1/ai, use your YepAPI key, and pass the NVIDIA model ID (for example "nvidia/nemotron-3.5-lightning").