Nemotron 3 Super API Prices Rise 433%: What Developers Need to Budget Now

The Numbers

NVIDIA's Nemotron 3 Super has received a steep pricing revision. Input costs moved from $0.08 to $0.30 per 1M tokens, and output costs climbed from $0.45 to $2.40 per 1M tokens — a blended increase that works out to roughly +433% across the board.

That's not a rounding error. Output tokens, which typically drive the bulk of costs in generation-heavy workloads, are now more than five times more expensive than before.

Does This Matter for Your Workload?

It depends heavily on your output-to-input ratio:

• Low-output tasks (classification, short answers, embeddings pipelines): The input jump from $0.08 to $0.30 stings, but remains in budget for moderate volumes.
• High-output tasks (summarization, code generation, long-form drafting): The output leap from $0.45 to $2.40 is material. A workload generating 10M output tokens monthly just went from $4.50 to $24.00 — monthly, before input costs.
• High-volume production apps: Rerun your cost models now. A +433% increase is large enough to flip ROI calculations on pipelines that were previously marginal.

What to Do Next

If Nemotron 3 Super is central to your stack, this is a good moment to benchmark alternatives. Several models in a similar capability tier have held steadier pricing recently. Check the Nemotron 3 Super — live specs & price history page for full context and side-by-side comparisons before committing to a new budget.

Price increases of this magnitude from infrastructure providers are worth tracking closely — they rarely reverse quickly.

Ezra, Scout AI Team