nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-GenRM

NVIDIA-Nemotron-3-Ultra-550B-A55B-GenRM


Model Summary

| | | |:---|:---| | Total Parameters | 550B (55B active) | | Architecture | LatentMoE - Mamba-2 + MoE + Attention hybrid with Multi-Token Prediction (MTP) | | Context Length | Up to 1M tokens | | Minimum GPU Requirement | 8x GB200/B200/GB300/B300, 16x H100, 8x H200 | | Supported Languages | English, French, Spanish, Italian, German, Japanese, Hindi, Korean, Brazilian Portuguese, and Chinese | | Best For | Judging Model Responses | | Reasoning Mode | Thinking On Only | | License | OpenMDW License Agreement, version 1.1 | | Release Date | June 4, 2026 |

Quick Start

For more details on how to deploy and use the model - see the [Quick Start Guide](#quick-start-guide) below!

Model Overview

Model Developer: NVIDIA Corporation

Model Dates: December 2025 - April 2026

Data Freshness:

• The post-training data has a cutoff date of May 2026.
• The pre-training data has a cutoff date of September 2025.

What is Nemotron?

NVIDIA Nemotron™ is a family of open models with open weights, training data, and recipes, delivering leading efficiency and accuracy for building specialized AI agents.

Description

NVIDIA-Nemotron-3-Ultra-550B-A55B-GenRM is a Generative Reward Model (GenRM) that leverages NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16 as the foundation and is fine-tuned to evaluate the quality of assistant's responses.

Given a conversation history, a new user request, and two candidate assistant responses, it produces an individual helpfulness score for each response and a ranking score. The model also accepts user-specified principles, when given the principles, it judges the responses based on the principles.

This GenRM is used in the Reinforcement Learning from Human Feedback training of NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16.

For training details, see the Nemotron 3 Ultra technical report.

This model is ready for commercial and non-commercial use.

License/Terms of Use

Use of this model is governed by the OpenMDW License Agreement, version 1.1 (OpenMDW-1.1).

Deployment Geography: Global

Use Case

This GenRM is used in the Reinforcement Learning from Human Feedback training of NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16.

Release Date

Hugging Face - 06/04/2026 via Hugging Face

Reference(s)

• NVIDIA Nemotron 3 model family on Hugging Face
• NVIDIA Nemotron 3 Ultra Technical Report

Model Architecture

• Architecture Type: Mamba2-Transformer Hybrid Latent Mixture of Experts (LatentMoE) with Multi-Token Prediction (MTP)
• Network Architecture: Nemotron Hybrid LatentMoE
• Number of model parameters: 550B Total / 55B Active

Model Design

We developed this model using an early version of NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16 as its foundation. This model contains 550 billion parameters.

Input

• Input Type(s): Text
• Input Format(s): String
• Input Parameters: One-Dimensional (1D): Sequences
• Other Properties Related to Input: Maximum context length up to 1M tokens. Supported languages include: English, French, Spanish, Italian, German, Japanese, Korean, Hindi, Korean, Brazilian Portuguese, and Chinese

Output

• Output Type(s): Text
• Output Format: String
• Output Parameters: One-Dimensional (1D): Sequences
• Other Properties Related to Output: Maximum context length up to 1M tokens

Our AI models are designed and optimized to run on NVIDIA GPU-accelerated systems. By leveraging NVIDIA's hardware (e.g. GPU cores) and software frameworks (e.g., CUDA libraries), the model achieves faster training and inference times compared to CPU-only solutions.

Software Integration

• Runtime Engine(s): NeMo 26.04.01
• Supported Hardware Microarchitecture Compatibility: NVIDIA Ampere - A100; NVIDIA Blackwell; NVIDIA Hopper - H100-80GB
• Operating System(s): Linux

The integration of foundation and fine-tuned models into AI systems requires additional testing using use-case-specific data to ensure safe and effective deployment. Following the V-model methodology, iterative testing and validation at both unit and system levels are essential to mitigate risks, meet technical and functional requirements, and ensure compliance with safety and ethical standards before deployment.

Model Version(s)

• v1.0 - GA

Quick Start Guide

The model shares the same architecture as NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16.

> Deployment instructions can be found on the NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16 model card.

Now you can query the model, here is an example:

from openai import OpenAI
client = OpenAI(base_url="http://127.0.0.1:8000/v1", api_key="dummy")
msg = [
{"role": "user", "content": "What is 1+1?"},
{"role": "assistant", "content": "1+1=2"},
{"role": "user", "content": "What about 1+2?"},
{"role": "response_1", "content": "1+2=4"},
{"role": "response_2", "content": "1+2=3"}
]
completion = client.chat.completions.create(
model="nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-GenRM",
messages=msg,
temperature=1.0,
top_p=0.95,
max_tokens=24576,
stream=False
)
output = completion.choices[0].message.content
print(output)

You can also use it with pre-defined principles by formatting message as:

msg = [
{"role": "user", "content": "How's the weather in LA?"},
{"role": "response_1", "content": "I don't have access to real-time data, so I can't give you the current weather in Los Angeles."},
{"role": "response_2", "content": "Most days sit in the 65 °F–80 °F (18 °C–27 °C) range, with cooler evenings, especially near the coast."},
{"role": "principle", "content": "You will be given one or more evaluation criteria…