Skip to main content

Nemotron Ultra API pricing, benchmarks and specs | AshnaAI

NVIDIA

Nemotron Ultra

Nemotron Ultra is NVIDIA's large mixture-of-experts model for difficult reasoning and multi-step agent tasks.

Try this model API docs

Model details

Model ID: nemotron-ultra
Provider: NVIDIA
Modality: Text generation
Context window: 202,000 tokens
Input → output: Text → text
Released: Jun 4, 2026
Available on AshnaAI: Oct 3, 2026

Where this model fits

Best capabilities

Tool calling and agent workflows
Web-connected research tasks

Best for

General chat, writing, summarization, and structured extraction
Tasks where response quality matters more than minimum latency
Agents that need to call APIs or execute multi-step workflows

Not ideal for

Image, screenshot, or visual-document analysis
Ultra-low-cost, latency-sensitive, high-volume classification

Benchmarks

No comparable, source-verified benchmark has been published for this catalog entry yet.

Pricing (USD)

Limited-time 30% discount

Input / 1M tokens: $0.723$0.506
Output / 1M tokens: $2.89$2.02
Cached input / 1M: $0.145$0.101
Cache write / 1M: —

Capabilities

reasoning—
toolsSupported
vision—
web searchSupported
files—

Available regions

US East (N. Virginia)

us-east-1

Europe (Frankfurt)

eu-central-1

Asia Pacific (Mumbai)

ap-south-1

Multi-region and provider failover: Active

Requests automatically switch to another region or provider if one becomes unavailable.