nvidia-nemotron-3-super-120b-a12b API price comparison
An open-source, efficient hybrid Mamba-Transformer MoE model that supports a context length of one million tokens and excels at agent reasoning, programming, planning, and tool invocation.
Cheapest overall
AIHubMix
0.66 USD/M tokens combined per 1M tokens
Lowest input
AIHubMix
0.11 USD/M tokens input per 1M tokens
Lowest output
AIHubMix
0.55 USD/M tokens output per 1M tokens
Last refreshed
Sep 4, 2026
1 listed platform
Best listed option
nvidia-nemotron-3-super-120b-a12b is currently cheapest overall on AIHubMix, with input at 0.11 USD/M tokens and output at 0.55 USD/M tokens per million tokens.
Pricing insight
Across 1 listed routes, nvidia-nemotron-3-super-120b-a12b is cheapest overall on AIHubMix at 0.66 USD/M tokens.
-
Lowest total cost
AIHubMix
Best first check for balanced coding-agent sessions: 0.66 USD/M tokens.
-
Prompt-heavy work
AIHubMix
Useful for large repositories and long context reads: 0.11 USD/M tokens input.
-
Response-heavy work
AIHubMix
Useful for long answers, edits, and agent loops: 0.55 USD/M tokens output.
Median combined price
0.66 USD/M tokens
Baseline for judging whether the cheapest route is unusually low.
Cache price coverage
100%
1 of 1 routes publish cache read or write prices.
Fresh rows
1/1
Routes crawled in the last 72 hours.
Use case fit
nvidia-nemotron-3-super-120b-a12b from Other can be compared by input, output, and combined per-million-token cost to select the most economical API route.
Platform price comparison
| Platform | Ratio | Input | Output | Total |
|---|---|---|---|---|
| AIHubMix | 1.00x | 0.11 USD/M tokens | 0.55 USD/M tokens | 0.66 USD/M tokens |
FAQ
Where is nvidia-nemotron-3-super-120b-a12b cheapest?
AIHubMix has the lowest listed combined price at 0.66 USD/M tokens per 1M tokens.
What do input and output prices mean?
Input is charged for prompt tokens you send. Output is charged for tokens generated by the model.
How often are prices refreshed?
Prices are crawler-backed and this page was last refreshed Sep 4, 2026. Always confirm before large purchases.