NVIDIA API pricing
As of October 22, 2026, the NVIDIA API prices run from $0.147 to $2.40 per million output tokens across 7 models, as listed on OpenRouter. Nemotron 3.5 Lightning is the cheapest paid model at $0.0534 input and $0.147 output per million tokens, and Nemotron 3 Ultra the most expensive at $0.60 input and $2.40 output.
Nemotron 3 Ultra ranks first among NVIDIA models in our Vibe Coding Index, at 30.0 on 3 benchmarks. 2 of 7 NVIDIA models have any benchmark score. Cost per solved task is known for 1 of 7 models.
| Model | Input $/Mtok | Output $/Mtok | Context | Evidence | Index | Cost per solved task |
|---|---|---|---|---|---|---|
| Nemotron 3.5 Super VL | $0.15 | $0.50 | 1.05M | Not enough data | No run cost | |
| Switchyard | Not published | Not published | 1M | Not benchmarked | No run cost | |
| Nemotron 3.5 Lightning | $0.0534 | $0.147 | 262K | 22.6 | $0.06 | |
| Nemotron 3 Ultra | $0.60 | $2.40 | 262K | 30.0 | No run cost | |
| Nemotron 3.5 Content Safety | $0.20 | $0.20 | 131K | Not benchmarked | No run cost | |
| Nemotron 3 Super | $0.085 | $0.45 | 262K | Not benchmarked | No run cost | |
| Nemotron 3 Nano 30B A3B | $0.055 | $0.22 | 262K | Not benchmarked | No run cost |
Price changes
No change in a maker's own API price recorded yet. We read prices daily from OpenRouter (since October 15, 2026) and list a change only when the maker's own price moves between two of our snapshots.