NvidiabalancedTool use

Nemotron 3 Ultra

NVIDIA Nemotron 3 Ultra is an open frontier-reasoning and orchestration model from NVIDIA, with 55B active parameters out of 550B total (MoE). Built on a hybrid Transformer-Mamba mixture-of-experts architecture, it...

1M context·~750K words·16K max output

View context details Compare with Jp Anthropic Claude Sonnet 4 6

Context window1Mtokens

Max output16Ktokens

Context window

This model accepts 1M tokens in one request (~750K words of text).

Context window size1M tokens

4K32K128K1M10M

What fits in one request

Short document
About 1,500 words of text
Fits
Long document
About 37K words of text
Fits
Small codebase
About 150K words of text
Fits
Full novel
About 375K words of text
Fits

Specifications

Context size, pricing, and release info in one place.

Context window: 1,000,000 tokens (1M)
Max output tokens: 16,384 tokens (16K)
Speed tier: balanced
Provider: Nvidia
Release date: Jun 2026

Capabilities

See which features this model supports, such as vision, tools, and streaming.

Supported (5)

Tool use: Can call external tools and APIs; Supported
Function calling: Structured function call interface; Supported
Extended thinking: Shows its chain-of-thought reasoning; Supported
Streaming: Returns tokens as they are generated; Supported
Prompt caching: Reuse repeated prompt prefixes cheaply; Supported

Not supported (3)

Vision: Accepts image inputs alongside text; Not supported
Web search: Can browse the web during a request; Not supported
Batch API: Process many requests asynchronously; Not supported

Best for

Jump to a guide or ranking that matches each workload.

Reasoning AI agents Long documents RAG

Compare Nemotron 3 Ultra

Open a side-by-side comparison with one click.

Frequently asked questions

Short answers about context size and how this model behaves.

Nemotron 3 Ultra has a context window of 1M tokens (1,000,000 tokens). This million-token window can process entire codebases, long legal documents, or book-length texts in a single pass.

More from Nvidia

Other models by Nvidia in our catalog.

Nemotron 3 Ultra (free)

1Mbalanced

Nemotron 3 Nano 30B A3B

262Kfast

Nemotron 3 Super

262Kbalanced

Nemotron 3 Super (free)

262Kbalanced

Nvidia Nemotron Nano 3 30b

262Kfast

Nemotron 3 Nano 30B A3B (free)

256Kfast

Use a smaller model.
Get better results.

Mem0 gives your AI long-term memory so you stop re-sending context on every call. That means you can use a smaller, faster, cheaper model — and still get better answers.

Calculate your savings Learn about Mem0

Example: a multi-turn chat session

Without Mem0~128K tokens sent

Full history

Repeated info

Old context

With Mem0~20K tokens sent

Key memories

Current turn

80% less to send — works with any model