NVIDIA API pricing by model
Updated
BroadRouter tracks API prices for 10 NVIDIA models, 5 of them listed by 2 or more providers. In our snapshot of 2026-11-14, the lowest listed input price is USD 0.05 per 1M tokens, for NVIDIA Nemotron 3 Nano 30B A3B at Novita AI. BroadRouter is BroadNet's site.
- Models priced
- 10
- Listed by 2+ providers
- 5
- Providers
- 16
- Lowest listed input / output
- USD 0.05 / USD 0.16
NVIDIA models and their lowest listed prices
The first 5 models have a standard price from 2 or more providers; the other 5 have one provider so far. USD per 1M tokens, standard tier. The input and output columns can come from different providers.
| Model | Providers | Lowest listed input | Lowest listed output | Maker's own API in / out | Checked |
|---|---|---|---|---|---|
| NVIDIA Nemotron 3 Ultra | 9 | USD 0.50 CoreWeave Forge | USD 2.15 CoreWeave Forge | — | 2026-11-15 |
| NVIDIA Nemotron 3.5 Super VL | 6 | USD 0.10 Crusoe Cloud | USD 0.40 Crusoe Cloud | — | 2026-11-15 |
| NVIDIA Nemotron 3.5 Lightning 30B A3B | 5 | USD 0.05 Crusoe Cloud | USD 0.16 DeepInfra | — | 2026-11-15 |
| NVIDIA Nemotron 3 Nano 30B A3B | 3 | USD 0.05 Novita AI | USD 0.20 Novita AI | — | 2026-11-15 |
| NVIDIA Nemotron 3 Super 120B A12B | 3 | USD 0.075 DeepInfra | USD 0.40 DeepInfra | — | 2026-11-14 |
| Nemotron 3 Nano Omni | 1 (single) | USD 0.30 Crusoe Cloud | USD 1.83 Crusoe Cloud | — | 2026-11-15 |
| Nemotron 3.5 Super | 1 (single) | USD 0.20 Perplexity | USD 0.50 Perplexity | — | 2026-11-15 |
| Nemotron-Content-Safety-3.5 | 1 (single) | USD 0.20 DeepInfra | USD 0.20 DeepInfra | — | 2026-11-14 |
| NVIDIA Nemotron Nano 2 | 1 (single) | USD 0.06 Amazon Bedrock | USD 0.23 Amazon Bedrock | — | 2026-11-14 |
| NVIDIA Nemotron Nano 2 VL | 1 (single) | USD 0.20 Amazon Bedrock | USD 0.60 Amazon Bedrock | — | 2026-11-14 |
"Lowest listed": the lowest standard, non-free price for the model in our snapshot, with the provider that lists it. Free rows and batch, flex, priority and fast tiers are on each model page but not in these columns. "Single": one provider so far; those pages are kept out of search results until a second provider lists the model. "Checked": the newest check date of the model's prices.
Where can you buy the NVIDIA API?
We have no price from an API run by NVIDIA in our data. 16 providers in our data list at least one NVIDIA model: DeepInfra (5), Amazon Bedrock (4), Crusoe Cloud (3), Venice (3), Baseten (2), CoreWeave Forge (2), Fireworks AI (2), Microsoft Azure (2), Cloudflare Workers AI (1), DigitalOcean (1), and 6 more.
Report a wrong price or number (email to info@broadnet.ai, with this page's link filled in).
All model makers · All models: API prices compared · Prices by provider · LLM API cost calculator
Questions
How many NVIDIA models does BroadRouter track?
10 NVIDIA models have a standard per-token price from at least one provider in our data, read from 16 providers' own pricing pages (newest check 2026-11-15). 5 of them are listed by 2 or more providers; the other 5 have one provider so far, and their pages are kept out of search results.
Which NVIDIA model has the lowest listed input price on 2026-11-14?
In our snapshot of 2026-11-14, NVIDIA Nemotron 3 Nano 30B A3B at Novita AI: USD 0.05 input and USD 0.20 output per 1M tokens, standard tier. The lowest listed output price is USD 0.16 per 1M tokens, for NVIDIA Nemotron 3.5 Lightning 30B A3B at DeepInfra (checked 2026-11-14). Free rows, and batch, flex and priority tiers, are left out. Price says nothing about quality, so compare models on your own task.
Does BroadNet serve any NVIDIA models?
No. BroadRouter is BroadNet's own site, and BroadNet Inference serves 4 chat models, none of them from NVIDIA (rate card synced 2026-11-14). The prices on this page are other providers' own prices. BroadNet Inference models.
Does NVIDIA sell its own API?
We have no price from an API run by NVIDIA in our data, so every price on this page is from a host.
Which providers list NVIDIA models?
16 providers publish a standard price for at least one NVIDIA model. By number of NVIDIA models priced: DeepInfra 5, Amazon Bedrock 4, Crusoe Cloud 3, Venice 3, Baseten 2, and 11 more. Each model page ranks its providers on a blend of 3 input tokens to 1 output token.