Nemotron 3 Ultra

By NVIDIA, United States

Very large model from NVIDIA, free through OpenRouter. Its provider may train its models on your messages.

May trainHosted in the USOpen modelFree

Where your message goes

A route is the provider that runs the model for aidock. The chat tries the main route first, then the others if it is busy, and tells you.

OpenRouter

Free

Main route

May trainHosted in the US

Training: Its provider may reuse what you write to train its models.
Retention: The final provider (NVIDIA) may reuse messages to train its models.
Data processing agreement: Not stated
Transfer outside the EU: Standard contractual clauses (OpenRouter is not listed under the Data Privacy Framework).
Provider’s free quota: ":free" models: 20 requests a minute, 50 a day, or 1,000 a day after a one-time purchase of $10 in credits.

Sources

Information checked on 11 October 2026 in the providers’ official documents. Technical words are explained on the Models page.

Key facts

Context: 1,000,000 tokens; How much text it reads at once (a token is about 4 letters).
Reads images: No
Thinks before answering: Yes
Open weights: Yes

Strengths

Reasoning
Long documents

Speed measured this week

Anonymous measurements on real requests made in the chat, over 7 days. No text is kept.

Loading measurements…

Quality of answers

aidock does not rate the quality of answers. To compare it, see compar:IA, the French government’s public comparison arena: anyone can compare two models blind and vote there.

Open compar:IA

Frequently asked questions

Is Nemotron 3 Ultra free?

Yes. In the aidock chat, Nemotron 3 Ultra is free: 10 messages a day without an account, 100 with a free account. aidock uses the free offer of OpenRouter.

Are my messages used to train Nemotron 3 Ultra?

It is possible. Its provider may reuse what you write to train its models. The chat warns you before your first message to this model.

Where is my data processed with Nemotron 3 Ultra?

OpenRouter: United States.

All models