Nemotron 3 Ultra API Prices Rise: Input Up 4%, Output Up 30%

What Changed

NVIDIA has increased the API pricing for Nemotron 3 Ultra. Input tokens moved from $0.60 to $0.625 per 1M tokens — a modest bump most teams won't notice. Output tokens are the real story: up from $2.40 to $3.13 per 1M tokens, a +30% increase.

Who Actually Feels This

The input change is largely cosmetic. The output jump is meaningful if your workload skews toward generation — think long-form summarization, code generation, or agentic pipelines where the model writes substantially more than it reads. A workload producing 10M output tokens monthly now costs an extra $7.30 compared to before. Scale that up and the delta becomes a line item worth reviewing.

For retrieval-augmented or classification tasks where output is short, the impact stays small.

What to Do

If Nemotron 3 Ultra is a primary model in your stack, benchmark your actual input-to-output token ratio before assuming the worst or best. Check Nemotron 3 Ultra — live specs & price history for up-to-date pricing and to compare against alternatives on the site — particularly models in a similar capability tier that haven't moved prices recently.

Price increases on capable models aren't unusual, but a +30% output hike without an announced capability change is worth scrutinizing before your next billing cycle.

Ezra, Scout AI Team