NVIDIA Nemotron 3 Nano 30B A3B API Prices Rise 20% Across Input and Output

What Changed

NVIDIA has bumped the API price for Nemotron 3 Nano 30B A3B by +20% across both token types:

• Input: $0.05 → $0.06 per 1M tokens
• Output: $0.20 → $0.24 per 1M tokens

In absolute terms these are still modest numbers, but a uniform 20% increase is worth pressure-testing against your actual usage.

Does It Matter for Your Workload?

For low-volume or experimental use, the delta is negligible — you'd need to burn through tens of millions of tokens before the increase registers meaningfully on an invoice. But for production pipelines running heavy output loads (think summarization, code generation, or agentic loops), output costs compound fast. At $0.24 per 1M output tokens, a pipeline generating 100M output tokens monthly now costs $24 instead of $20 — not catastrophic, but a real line item.

If your workload is input-heavy (classification, embedding-adjacent tasks, retrieval-augmented generation with short responses), the input-side move from $0.05 to $0.06 is easier to absorb.

What to Do Next

Before adjusting budgets or switching models, benchmark actual token ratios from your logs — most real workloads skew output-heavy, which is where this increase bites hardest. Check the Nemotron 3 Nano 30B A3B — live specs & price history page for up-to-date pricing and to compare against alternatives in the same efficiency tier.

If Nemotron 3 Nano was on your shortlist primarily because of its low cost, this 20% rise makes a fresh comparison run worthwhile.

Ezra, Scout AI Team