Instructions to use nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-NVFP4 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-NVFP4 with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-NVFP4")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-NVFP4", dtype="auto")

Inference
HuggingChat
Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-NVFP4 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-NVFP4"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-NVFP4",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-NVFP4

SGLang

How to use nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-NVFP4 with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-NVFP4" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-NVFP4",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-NVFP4" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-NVFP4",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-NVFP4 with Docker Model Runner:
```
docker model run hf.co/nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-NVFP4
```

Configuration Parsing Warning:Config file config.json cannot be fetched (too big)

NVIDIA-Nemotron-3-Ultra-550B-A55B-NVFP4

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	4xGB200, 4xB200, 4x GB300, 4x B300, 8xH100
Supported Languages	English, French, Spanish, Italian, German, Japanese, Korean, Hindi, Korean, Brazilian Portuguese, and Chinese
Best For	Frontier reasoning, complex agentic workflows, long-context analysis, tool use, multilingual reasoning, high-stakes RAG
Reasoning Mode	Configurable on/off via chat template (`enable_thinking=True/False`)
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 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

Nemotron-3-Ultra-550B-A55B-NVFP4 is a frontier-scale large language model (LLM) trained by NVIDIA, designed to deliver strong agentic, reasoning, and conversational capabilities. It is optimized for the most demanding workloads, including complex multi-step agents, long-context analysis, and high-accuracy reasoning over code, math, and science. Like other models in the family, it responds to user queries and tasks by first generating a reasoning trace and then concluding with a final response. The model's reasoning capabilities can be configured through a flag in the chat template.

The model employs a hybrid Latent Mixture-of-Experts (LatentMoE) architecture, utilizing interleaved Mamba-2 and MoE layers, along with select Attention layers. Like the Super model, the Ultra model incorporates Multi-Token Prediction (MTP) layers for faster text generation and improved quality, and it is trained using an NVFP4 pre-training recipe to maximize compute efficiency. The model has 55B active parameters and 550B parameters in total.

The supported languages include: English, French, Spanish, Italian, German, Japanese, Korean, Hindi, Korean, Brazilian Portuguese, and Chinese.

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

License/Terms of Use

Governing Download Terms: Use of this model is governed by the OpenMDW-1.1 model license.

Benchmarks

Benchmark	Nemotron 3 Ultra BF16	Nemotron 3 Ultra NVFP4
Agentic
Terminal Bench 2.1	56.4	53.9
GDPVal	46.7	47.9
SWE-Bench Verified	71.9	69.7
SWE-Bench Multilingual	67.7	65.8
ProfBench (Search)	56	56.4
PinchBench	90	89.8
TauBench V3
Airline	81.5	80.0
Retail	86.4	88.4
Telecom	92.9	93.6
Banking	22.6	19.2
Average	70.9	70.3
BrowseComp	44.4	41.4
Reasoning and Knowledge
IOI 2025	570.0	564.7
GPQA (no tools)	87.0	87.9
SciCode (subtask)	44.6	43.5
HLE (no tools)	26.7	26.1
CritPt (no tools)	3.1	3.4
OmniScience Accuracy	24.1	24.6
OmniScience Non-Hallucination	78.7	75.5
Chat & Instruction Following
IFBench (prompt)	81.7	82.3
Long Context
AA-LCR	65.4	65.5
RULER 1M	94.7	94.0

All evaluation results were collected via Nemo Evaluator SDK. We used three main evaluation harnesses: Nemo Gym, Nemo Skills, and Harbor with extended sandboxing support via AWS ECS on Nemo Evaluator. In addition, the evaluations also used dedicated open-source packaged containers for ScaleAI Multi Challenge Multi Turn Instruction Following and KernelBench. For reproducibility purposes, more details on the evaluation settings and pinned containers can be found in the Nemo Evaluator SDK examples folder and the reproducibility tutorial for Nemotron 3 Ultra.

The following benchmarks are not onboarded yet in our open source tools and for these we used either their official open source implementation or otherwise an internal scaffolding that we plan to open source in the future: BrowseComp with Search, Tau Bench 3, ProfBench with Search, PinchBench, Vals.ai, LongBench v2.

Deployment Geography: Global

Use Case

NVIDIA-Nemotron-3-Ultra-550B-A55B-NVFP4 is a frontier-scale general purpose reasoning and chat model intended to be used in English, Code, and supported multilingual contexts. This model is optimized for complex agentic workflows, long-context reasoning, and high-stakes analytical workloads. It is intended to be used by developers designing AI Agent systems, chatbots, RAG systems, and other AI-powered applications. This model is also suitable for complex instruction-following tasks and long-context reasoning over very large documents and codebases.

Release Date

Hugging Face - 06/04/2026 via Hugging Face

Reference(s)

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

The model utilizes the LatentMoE architecture, where tokens are projected into a smaller latent dimension for expert routing and computation, improving accuracy per byte. The Ultra model is pre-trained using an NVFP4 recipe — sharing the quantization-aware pre-training approach pioneered in the Nemotron 3 family. The majority of linear layers use NVFP4 for weights, activations, and gradients, while select layers (including latent projections, MTP layers, QKV/attention projections, and embeddings) are maintained in BF16 or MXFP8 for training stability. The model includes Multi-Token Prediction (MTP) layers using a shared-weight design across prediction heads. This improves training signal quality, enables faster inference via native speculative decoding, and supports more stable autoregressive drafting at longer draft lengths compared to independently trained offset heads.

Training Methodology

Stage 1: Pre-Training

NVIDIA-Nemotron-3-Ultra-550B-A55B-Base-BF16 model was pre-trained for approximately 20T tokens using crawled and synthetic code, math, science, and general knowledge data. Training leveraged an NVFP4 recipe for efficiency. All datasets are disclosed in the Training and Evaluation Datasets section of this document. Major portions of the pre-training corpus are released in the Nemotron-Pre-Training-Datasets collection.
Software used for pre-training: Megatron-LM

Stage 2: Supervised Fine-Tuning

The model was further fine-tuned on synthetic code, math, science, tool calling, instruction following, structured outputs, and general knowledge data. This stage incorporated data designed to support long-range retrieval and multi-document aggregation. All datasets are disclosed in the Training and Evaluation Datasets section of this document. Major portions of the fine-tuning corpus are released in the Nemotron-Post-Training-v3 collection. Data Designer is one of the libraries used to prepare these corpora.

Stage 3: Reinforcement Learning

The model underwent multi-environment reinforcement learning using asynchronous GRPO (Group Relative Policy Optimization) across math, code, science, instruction following, multi-step tool use, multi-turn conversations, and structured output environments. It utilized an asynchronous RL architecture that fully decouples training from inference across separate GPU devices, leveraging in-flight weight updates and MTP to accelerate rollout generation. Conversational quality was further refined through RLHF. All datasets are disclosed in the Training and Evaluation Datasets section of this document. The RL environments and datasets are released as part of NeMo Gym.
Software used for reinforcement learning: NeMo RL, NeMo Gym

Stage 4: Multi-Domain On-Policy Distillation (MOPD)

The model underwent Multi-Domain On-Policy Distillation (MOPD) to improve reasoning across many task types while staying efficient. This technique uses strong teacher models to guide training on the model's own generated attempts (on-policy rollouts), helping recover accuracy and improve performance across coding, math, instruction following, tool use, and agentic workflows. By distilling teacher signal onto the student's own trajectories rather than offline traces, MOPD better aligns the student's behavior with what it would actually produce at inference time, yielding stronger gains than purely off-policy distillation.

NVIDIA-Nemotron-3-Ultra-550B-A55B-NVFP4 model is a result of the above work.

The end-to-end training recipe is available in the NVIDIA Nemotron Developer Repository. Evaluation results can be replicated using the NeMo Evaluator SDK. Data Designer is one of the libraries used to prepare the pre and post training datasets. More details on the datasets and synthetic data generation methods can be found in the technical report NVIDIA Nemotron 3 Ultra Technical Report.

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 Ultra NVFP4 checkpoint is a frontier-scale model quantized for maximum throughput on the latest hardware. The minimum recommended hardware is:

Single-node: 4× B200 (fits NVFP4 weights plus KV cache with headroom)
Multi-node: ≥4 GPUs across GB200 / GB300

All deployment snippets below default to port 8000, with chunked prefill, NVFP4 KV caching, and MTP (5 speculative tokens) enabled.

Multi-Node Setup with Ray (Recommended for multi-node deployments)

The recommended multi-processing backend for multi-node deployments is Ray v2. Below is a template for launching a Ray cluster:

# Set the IP for the head node in RAY_HEAD_IP
export RAY_HEAD_IP=<head_node_ip>
export RAY_PORT=6379
export RAY_ADDRESS=${RAY_HEAD_IP}:${RAY_PORT}

# Start Ray head node (vLLM/SGLang will run on this node)
ray start --head --node-ip-address=${RAY_HEAD_IP} --port=${RAY_PORT}

# Start Ray worker node(s)
ray start --address=${RAY_HEAD_IP}:${RAY_PORT} --block

# Verify Ray cluster is ready
ray status --address=${RAY_HEAD_IP}:${RAY_PORT}

Note: ray[cgraph] is required: uv pip install "ray[cgraph]"

vLLM

Recommended container: vllm/vllm-openai:v0.22.0

For more detailed information, please see this cookbook.

export MODEL_CKPT=nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-NVFP4

4× B200 single-node deployment5:

docker run -d --name nemotron-ultra-vllm \
  --gpus all \
  --ipc=host \
  --network=host \
  --shm-size=16g \
  --ulimit memlock=-1 \
  --ulimit stack=67108864 \
  -v $MODEL_CKPT:/model:ro \
  -e VLLM_WORKER_MULTIPROC_METHOD=spawn \
  -e SAFETENSORS_FAST_GPU=1 \
  -e NVIDIA_TF32_OVERRIDE=1 \
  -e VLLM_LOGGING_LEVEL=INFO \
  vllm/vllm-openai:v0.22.0 \
  /model \
  --host 0.0.0.0 \
  --port 8000 \
  --served-model-name nvidia/nemotron-3-ultra \
  --trust-remote-code \
  --tensor-parallel-size 4 \
  --enable-expert-parallel \
  --kv-cache-dtype fp8 \
  --max-model-len 262144 \
  --gpu-memory-utilization 0.90 \
  --max-num-seqs 16 \
  --max-num-batched-tokens 32768 \
  --enable-chunked-prefill \
  --enable-prefix-caching \
  --reasoning-parser nemotron_v3 \
  --enable-auto-tool-choice \
  --tool-call-parser qwen3_coder \
  --mamba-ssm-cache-dtype float16 \
  --mamba-backend flashinfer \
  --enable-mamba-cache-stochastic-rounding \
  --mamba-cache-philox-rounds 5 \
  --speculative-config '{"method": "nemotron_h_mtp", "num_speculative_tokens": 5}' \
  --model-loader-extra-config '{"enable_multithread_load": true, "num_threads": 96}'

Multi-node deployment (e.g. 2× 4×GB300 with Ray): After launching the Ray head and worker per the multi-node setup above:

# Run on Ray head node
vllm serve $MODEL_CKPT \
  --host 0.0.0.0 \
  --port 8000 \
  --served-model-name nvidia/nemotron-3-ultra \
  --tensor-parallel-size 4 \
  --pipeline-parallel-size 2 \
  --distributed-executor-backend ray \
  --trust-remote-code \
  --kv-cache-dtype fp8 \
  --gpu-memory-utilization 0.90 \
  --max-model-len 262144 \
  --max-num-seqs 256 \
  --max-num-batched-tokens 32768 \
  --enable-chunked-prefill \
  --enable-prefix-caching \
  --reasoning-parser nemotron_v3 \
  --enable-auto-tool-choice \
  --tool-call-parser qwen3_coder \
  --mamba-ssm-cache-dtype float16 \
  --mamba-backend flashinfer \
  --enable-mamba-cache-stochastic-rounding \
  --mamba-cache-philox-rounds 5 \
  --speculative-config '{"method": "nemotron_h_mtp", "num_speculative_tokens": 5}' \
  --model-loader-extra-config '{"enable_multithread_load": true, "num_threads": 96}' \
  --compilation-config '{"pass_config": {"fuse_allreduce_rms": false}}' \
  --distributed-timeout-seconds 3600

Context Length: Defaults to 256k above. To use up to 1M, set VLLM_ALLOW_LONG_MAX_MODEL_LEN=1 and --max-model-len 1048576.
Useful environment variables: VLLM_FLASHINFER_ALLREDUCE_BACKEND=trtllm, VLLM_FLASHINFER_MOE_BACKEND=latency (TRTLLM-Gen) or VLLM_FLASHINFER_MOE_BACKEND=throughput (CUTLASS).

SGLang

Container (tested on 4× B200): docker pull lmsysorg/sglang:v0.5.12.post1

For more detailed information, please see this cookbook.

4× B200 single-node deployment (NVFP4):

docker run -d --name nemotron-ultra-sglang \
  --gpus all \
  --cap-add SYS_NICE \
  --ipc=host \
  --network=host \
  --shm-size=16g \
  --ulimit memlock=-1 \
  --ulimit stack=67108864 \
  -v $MODEL_CKPT:/model:ro \
  -e SAFETENSORS_FAST_GPU=1 \
  -e NVIDIA_TF32_OVERRIDE=1 \
  -e SGLANG_DISABLE_DEEP_GEMM=1 \
  lmsysorg/sglang:v0.5.12.post1 \
  python3 -m sglang.launch_server \
  --model-path /model \
  --host 0.0.0.0 \
  --port 8000 \
  --served-model-name nvidia/nemotron-3-ultra \
  --tp-size 4 \
  --ep-size 4 \
  --context-length 262144 \
  --mem-fraction-static 0.85 \
  --chunked-prefill-size 32768 \
  --fp8-gemm-backend triton \
  --moe-runner-backend triton \
  --mamba-scheduler-strategy no_buffer \
  --disable-piecewise-cuda-graph \
  --reasoning-parser nemotron_3 \
  --tool-call-parser qwen3_coder \
  --speculative-algorithm EAGLE \
  --speculative-num-steps 5 \
  --speculative-eagle-topk 1 \
  --speculative-num-draft-tokens 5 \
  --kv-cache-dtype fp8 \
  --trust-remote-code \
  --log-level info

Context length: Defaults to 256k above. To use up to 1M, set SGLANG_ALLOW_OVERWRITE_LONGER_CONTEXT_LEN=1 and --context-length 1048576.
Tool calls + reasoning parsing: When calling the chat completions endpoint with tools, you must set "chat_template_kwargs": {"enable_thinking": true, "force_nonempty_content": true} in the request body to parse both reasoning and tool calls correctly.

TRT-LLM

Container: docker pull nvcr.io/nvidia/tensorrt-llm/release:1.3.0rc17

For more detailed information, please see this cookbook.

Max-throughput configuration (B200×4):

cat > ./extra-llm-api-config.yml << EOF
cuda_graph_config:
  enable_padding: true
  max_batch_size: 16

enable_chunked_prefill: true
enable_attention_dp: true
max_num_tokens: 16384
num_postprocess_workers: 4
stream_interval: 10

kv_cache_config:
  dtype: fp8
  enable_block_reuse: false
  free_gpu_memory_fraction: 0.7
  mamba_ssm_cache_dtype: float16
  mamba_ssm_philox_rounds: 5
  mamba_ssm_stochastic_rounding: true
  mamba_state_cache_interval: 16384

moe_config:
  backend: CUTEDSL
  max_num_tokens: 65536
  use_low_precision_moe_combine: true

speculative_config:
  decoding_type: MTP
  max_draft_len: 3
  max_total_draft_tokens: 3
  num_nextn_predict_layers: 3
  speculative_model: null
  allow_advanced_sampling: true
EOF

trtllm-serve $MODEL_CKPT \
  --backend pytorch \
  --host 0.0.0.0 \
  --port 8000 \
  --max_batch_size 16 \
  --tp_size 4 --ep_size 4 \
  --max_num_tokens 16384 \
  --trust_remote_code \
  --reasoning_parser nano-v3 \
  --tool_parser qwen3_coder \
  --chat_template $MODEL_CKPT/chat_template.jinja \
  --extra_llm_api_options extra-llm-api-config.yml

Min-latency configuration (B200×4, NVFP4):

cat > ./extra-llm-api-config.yml << EOF
backend: pytorch
trust_remote_code: true
tensor_parallel_size: 4
pipeline_parallel_size: 1
context_parallel_size: 1
gpus_per_node: 8
moe_expert_parallel_size: 1
disable_overlap_scheduler: false

cuda_graph_config:
  batch_sizes:
  max_batch_size: 8
  enable_padding: true

enable_chunked_prefill: true
enable_attention_dp: false
max_batch_size: 8
max_seq_len: null
max_num_tokens: 8192
num_postprocess_workers: 0

kv_cache_config:
  dtype: fp8
  free_gpu_memory_fraction: 0.75
  mamba_state_cache_interval: 8192

moe_config:
  backend: TRTLLM

speculative_config:
  decoding_type: MTP
  max_draft_len: 5
  max_total_draft_tokens: 5
  # rc14 compatibility: rc14 derives max_draft_len from this old field.
  # rc15+ keeps max_draft_len and ignores this as deprecated.
  num_nextn_predict_layers: 5
  allow_advanced_sampling: true
EOF

trtllm-serve  \
<nvfp4_ckpt> \
--max_batch_size 8 \
--tp_size 4 --ep_size 1 \
--max_num_tokens 16384 \
--trust_remote_code \
--reasoning_parser nano-v3 \
--tool_parser qwen3_coder \
--chat_template $MODEL_DIR/chat_template.jinja \
--extra_llm_api_options extra-llm-api-config.yml

Long-context configuration: For long-context benchmarking, set TLLM_ALLOW_LONG_MAX_MODEL_LEN=1 as an environment variable and add --max_seq_len <seq_len> as the desired maximum context length. The MTP speculative_config block above carries over unchanged — on rc16, max_draft_len is the authoritative field and num_nextn_predict_layers is treated as deprecated.

API Client

The examples below use the OpenAI-compatible client and work with any of the serving backends above.

NOTE: For coding agents add the following to the API call - extra_body={"chat_template_kwargs": {"force_nonempty_content": True}}

from openai import OpenAI
client = OpenAI(base_url="http://localhost:8000/v1", api_key="EMPTY")
MODEL = "nvidia/nemotron-3-ultra"

Reasoning ON (default)

response = client.chat.completions.create(
    model=MODEL,
    messages=[{"role": "user", "content": "Write a haiku about GPUs"}],
    max_tokens=16000,
    temperature=1.0,
    top_p=0.95,
    extra_body={"chat_template_kwargs": {"enable_thinking": True}}
)
print(response.choices.message.content)

Reasoning OFF

response = client.chat.completions.create(
    model=MODEL,
    messages=[{"role": "user", "content": "What is the capital of Japan?"}],
    max_tokens=16000,
    temperature=1.0,
    top_p=0.95,
    extra_body={"chat_template_kwargs": {"enable_thinking": False}}
)
print(response.choices.message.content)

Medium-effort reasoning

Uses significantly fewer reasoning tokens than full thinking mode. Recommended as a starting point before tuning explicit token budgets.

response = client.chat.completions.create(
    model=MODEL,
    messages=[{"role": "user", "content": "What is the capital of Japan?"}],
    max_tokens=16000,
    temperature=1.0,
    top_p=0.95,
    extra_body={"chat_template_kwargs": {"enable_thinking": True, "medium_effort": True}}
)
print(response.choices[0].message.content)

Tool calling with reasoning (SGLang requires explicit chat template kwargs)

response = client.chat.completions.create(
    model=MODEL,
    messages=[{"role": "user", "content": "What's the weather in New York?"}],
    tools=[{
        "type": "function",
        "function": {
            "name": "get_weather",
            "description": "Get the current weather for a city.",
            "parameters": {
                "type": "object",
                "properties": {
                    "city": {"type": "string"},
                    "unit": {"type": "string", "enum": ["celsius", "fahrenheit"]}
                },
                "required": ["city"]
            }
        }
    }],
    tool_choice="required",
    max_tokens=256,
    temperature=1.0,
    top_p=0.95,
    extra_body={"chat_template_kwargs": {"enable_thinking": True, "force_nonempty_content": True}}
)

OpenCode

OpenCode is an AI coding agent that runs in your terminal. It connects to any OpenAI-compatible endpoint, making it compatible with all three serving backends above (vLLM, SGLang, and TRT-LLM).

Create or update your ~/.config/opencode/opencode.json:

{
    "$schema": "https://opencode.ai/config.json",
    "model": "local/nvidia-nemotron-3-ultra",
    "provider": {
        "local": {
            "npm": "@ai-sdk/openai-compatible",
            "name": "local_backend",
            "options": {
                "baseURL": "http://localhost:8000/v1",
                "apiKey": "EMPTY"
            },
            "models": {
                "nvidia-nemotron-3-ultra": {
                    "name": "nvidia/nemotron-3-ultra",
                    "limit": {
                        "context": 1000000,
                        "output": 32768
                    }
                }
            }
        }
    },
    "agent": {
        "build": {
            "temperature": 1.0,
            "top_p": 0.95,
            "max_tokens": 32000
        },
        "plan": {
            "temperature": 1.0,
            "top_p": 0.95,
            "max_tokens": 32000
        }
    }
}

All backends above default to port 8000, so the baseURL works as-is for vLLM, SGLang, and TRT-LLM. To learn more about other supported agent scaffolds, check out this resource.

Set a hard token ceiling on the reasoning trace using reasoning_budget. The model will attempt to close the trace at the next newline before the budget is hit; if none is found within 500 tokens it closes abruptly at reasoning_budget + 500.

from typing import Any, Dict, List
import openai
from transformers import AutoTokenizer


class ThinkingBudgetClient:
    def __init__(self, base_url: str, api_key: str, tokenizer_name_or_path: str):
        self.tokenizer = AutoTokenizer.from_pretrained(tokenizer_name_or_path)
        self.client = openai.OpenAI(base_url=base_url, api_key=api_key)

    def chat_completion(
        self,
        model: str,
        messages: List[Dict[str, Any]],
        reasoning_budget: int = 512,
        max_tokens: int = 1024,
        **kwargs,
    ) -> Dict[str, Any]:
        assert max_tokens > reasoning_budget, (
            f"reasoning_budget must be less than max_tokens. "
            f"Got {max_tokens=} and {reasoning_budget=}"
        )

        # Step 1: generate the reasoning trace up to the budget
        response = self.client.chat.completions.create(
            model=model, messages=messages, max_tokens=reasoning_budget, **kwargs
        )
        reasoning_content = response.choices.message.content
        if "</think>" not in reasoning_content:
            reasoning_content = f"{reasoning_content}.\n\n</think>\n"

        reasoning_tokens_len = len(
            self.tokenizer.encode(reasoning_content, add_special_tokens=False)
        )
        remaining_tokens = max_tokens - reasoning_tokens_len
        assert remaining_tokens > 0, (
            f"No tokens remaining for response ({remaining_tokens=}). "
            "Increase max_tokens or lower reasoning_budget."
        )

        # Step 2: continue from the closed reasoning trace
        messages.append({"role": "assistant", "content": reasoning_content})
        prompt = self.tokenizer.apply_chat_template(
            messages, tokenize=False, continue_final_message=True
        )
        response = self.client.completions.create(
            model=model, prompt=prompt, max_tokens=remaining_tokens, **kwargs
        )

        return {
            "reasoning_content": reasoning_content.strip().strip("</think>").strip(),
            "content": response.choices.text,
            "finish_reason": response.choices.finish_reason,
        }

# Example usage (32-token reasoning budget):
client = ThinkingBudgetClient(
    base_url="http://localhost:8000/v1",
    api_key="EMPTY",
    tokenizer_name_or_path="nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-NVFP4",
)

result = client.chat_completion(
    model="nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-NVFP4",
    messages=[
        {"role": "system", "content": "You are a helpful assistant. /think"},
        {"role": "user", "content": "What is 2+2?"},
    ],
    reasoning_budget=32,
    max_tokens=512,
    temperature=1.0,
    top_p=0.95,
)
print(result)

Training and Evaluation Datasets

Training

Data Modality: Text
The total size: 53.8 TiB (14.8 trillion tokens)
Total number of datasets: 226
Dataset partition: Training [100%], testing [0%], validation [0%]
Time period for training data collection: 2013 to 2026
Time period for testing data collection: 2013 to 2026
Time period for validation data collection: 2013 to 2026
Data Collection Method by dataset: Hybrid: Automated, Human, Synthetic
Labeling Method by dataset: Hybrid: Automated, Human, Synthetic

NVIDIA-Nemotron-3-Ultra-550B-A55B-NVFP4 is pre-trained on a large corpus of high-quality curated and synthetically-generated data. It is trained in the English language, as well as 11 other languages and 43 programming languages. Our sources cover a variety of document types such as: webpages, dialogue, articles, and other written materials. The corpus spans domains including legal, math, science, finance, and more. We also include a small portion of question-answering, and alignment style data to improve model accuracy. The model was pre-trained for approximately 20 trillion tokens.

The post-training corpus for NVIDIA-Nemotron-3-Ultra-550B-A55B-NVFP4 consists of high-quality curated and synthetically-generated data. Primary languages used for post-training include English, French, Spanish, Italian, German, Japanese, Korean, Hindi, Korean, Brazilian Portuguese, and Chinese

These datasets, such as FinePDFs, EssentialWeb, HotpotQA, SQuAD, and HelpSteer3, do not collectively or exhaustively represent all demographic groups (and proportionally therein). For instance, these datasets do not contain explicit mentions of demographic classes such as age, gender, or ethnicity in 64-99% of samples, depending on the source. In the subset where such terms are present, document-based datasets (FinePDFs and EssentialWeb) contain representational skews, such as references to "male" outnumbering those to "female", and mentions of "White" as the most frequent among ethnic identifiers (comprising 43-44% of ethnicity mentions). To mitigate these imbalances, we recommend considering evaluation techniques such as bias audits, fine-tuning with demographically balanced datasets, and mitigation strategies like counterfactual data augmentation to align with the desired model behavior. This evaluation used a 3,000-sample subset per dataset, identified as the optimal threshold for maximizing embedder accuracy.

During post-training, we generate synthetic data by distilling trajectories, solutions, and translations from strong teacher models and agent systems, often grounded in real tasks or documents and aggressively filtered for quality. For math, code, and science, we start from curated problem sets and use open source permissive models such as GPT-OSS-120B to produce step-by-step reasoning traces, candidate solutions, best-of-n selection traces, and verified CUDA kernels. For long-context and science, we build synthetic QA and reasoning data by retrieving passages from long documents, generating MCQ/OpenQA questions and answers, and paraphrasing them into multiple prompt/response formats to ensure diversity. Across all pipelines we stack automated verification—compilers, numerical checks, language identification—to ensure our data is high quality.

For all domains, we apply a unified data filtering pipeline to ensure that only high-quality, license-compliant, and verifiable samples are used for post-training. We first discard malformed examples using structural checks (e.g., missing tool definitions when tool calls are present). We then aggressively filter reasoning traces exhibiting pathological repetition, such as repeated n-grams within a sliding window or across the entire trajectory, which we found to be a strong indicator of malformed or low-quality reasoning. Finally, based on internal audits of synthetically generated datasets, we observed that some teacher models occasionally produce reasoning traces and final responses that implicitly align with specific political entities or promote nationalistic narratives. To mitigate this, we apply targeted keyword- and regex-based filters and remove all trajectories matching such behavior.

Alongside the model, we release our final pre-training and post-training data, as outlined in this section. For ease of analysis, there is a sample set that is ungated. For all remaining code, math and multilingual data, gating and approval is required, and the dataset is permissively licensed for model training purposes.

More details on the datasets and synthetic data generation methods can be found in the technical report NVIDIA Nemotron 3 Ultra.

For Detailed Dataset Information: Click here!

Base Pre-Training Corpus (Nemotron 3 Foundation)

The foundation of the model is trained on the Nemotron-3-Ultra corpus, comprising the following datasets from the Nemotron Pretraining Datasets collection:

Dataset Collection	Token Counts	Description
Nemotron-CC-v2 & v2.1	9.1T	A massive collection of English web data filtered from Common Crawl, including 2.5T+ tokens of new organic, translated, and synthetically rephrased content.
Nemotron-CC-Code-v1	427.9B	High-quality code tokens extracted from Common Crawl using the Lynx + LLM pipeline to preserve structure and equations.
Nemotron-Pretraining-Code-v1 & v2 & v3	1.7T	Curated GitHub code references with multi-stage filtering, deduplication, and large-scale synthetic code data.
Nemotron-CC-Math-v1	133.3B	High-quality math pre-training dataset preserving LaTeX formatting and mathematical structures.
Nemotron-Pretraining-Specialized-v1 & v1.1 & v1.2 & Nemotron-Pretraining-SFT-v1	660.0B	Synthetic datasets targeting specialized domains such as STEM reasoning and scientific coding.
Nemotron-Pretraining-Legal-v1	4.3B	Synthetic datasets targeting the legal domain.

Public Datasets

Dataset	Collection Period
GSM8K	4/23/2025
CC-NEWS	4/23/2025
Common Crawl	4/23/2025
Wikimedia	4/23/2025
Bespoke-Stratos-17k	4/23/2025
tigerbot-kaggle-leetcodesolutions-en-2k	4/23/2025
glaive-function-calling-v2	4/23/2025
APIGen Function-Calling	4/23/2025
LMSYS-Chat-1M	4/23/2025
Open Textbook Library - CC BY-SA & GNU subset and OpenStax - CC BY-SA subset	4/23/2025
Advanced Reasoning Benchmark, tigerbot-kaggle-leetcodesolutions-en-2k, PRM800K, and SciBench	4/23/2025
FineWeb-2	4/23/2025
Court Listener	Legacy Download
peS2o	Legacy Download
OpenWebMath	Legacy Download
BioRxiv	Legacy Download
PMC Open Access Subset	Legacy Download
OpenWebText2	Legacy Download
Stack Exchange Data Dump	Legacy Download
PubMed Abstracts	Legacy Download
NIH ExPorter	Legacy Download
arXiv	Legacy Download
BigScience Workshop Datasets	Legacy Download
Reddit Dataset	Legacy Download
SEC's Electronic Data Gathering, Analysis, and Retrieval (EDGAR)	Legacy Download
Advanced Mathematical Problem Solving	Legacy Download
MathPile	Legacy Download
NuminaMath CoT	Legacy Download
PMC Article	Legacy Download
FLAN	Legacy Download
Advanced Reasoning Benchmark	Legacy Download
SciBench	Legacy Download
WikiTableQuestions	Legacy Download
FinQA	Legacy Download
Riddles	Legacy Download
Problems in Elementary Mathematics for Home Study	Legacy Download
MedMCQA	Legacy Download
Cosmos QA	Legacy Download
MCTest	Legacy Download
AI2's Reasoning Challenge	Legacy Download
OpenBookQA	Legacy Download
MMLU Auxiliary Train	Legacy Download
social-chemestry-101	Legacy Download
Moral Stories	Legacy Download
The Common Pile v0.1	Legacy Download
FineMath	Legacy Download
MegaMath	Legacy Download
MegaMath	Legacy Download
MultiverseMathHard	10/2/2025
News Commentary	10/2/2025
Essential-Web	10/2/2025
finepdfs	10/2/2025
HotpotQA	10/2/2025
SQuAD2.0	10/2/2025
NLTK Words Lists	10/2/2025

Crawled and Scraped from Online Sources by NVIDIA

The English Common Crawl data was downloaded from the Common Crawl Foundation (see their FAQ for details on their crawling) and includes the snapshots CC-MAIN-2013-20 through CC-MAIN-2025-13. The data was subsequently deduplicated and filtered in various ways described in the Nemotron-CC paper. Additionally, we extracted data for fifteen languages from the following three Common Crawl snapshots: CC-MAIN-2024-51, CC-MAIN-2025-08, CC-MAIN-2025-18. The fifteen languages included were Arabic, Chinese, Danish, Dutch, French, German, Italian, Japanese, Korean, Polish, Portuguese, Russian, Spanish, Swedish, and Thai. As we did not have reliable multilingual model-based quality classifiers available, we applied just heuristic filtering instead—similar to what we did for lower quality English data in the Nemotron-CC pipeline, but selectively removing some filters for some languages that did not work well. Deduplication was done in the same way as for Nemotron-CC.

The GitHub Crawl was collected using the GitHub REST API and the Amazon S3 API. Each crawl was operated in accordance with the rate limits set by its respective source, either GitHub or S3. We collect raw source code and subsequently remove any having a license which does not exist in our permissive-license set (for additional details, refer to the technical report).

Dataset	Modality	Dataset Size	Collection Period	Collecting Organisation
English Common Crawl	Text	3.36T	4/8/2025	NVIDIA Advanced Deep Learning Research
English Common Crawl 1.1	Text	Not disclosed	10/2/2025	NVIDIA Advanced Deep Learning Research
Multilingual Common Crawl	Text	812.7B	5/1/2025	NVIDIA Advanced Deep Learning Research
GitHub Crawl	Text	747.4B	4/29/2025	NVIDIA Advanced Deep Learning Research
GitHub Crawl 1.1	Text	172.7B	9/30/2025	NVIDIA Advanced Deep Learning Research

Private Non-publicly Accessible Datasets of Third Parties

Dataset	Model(s) used
Global Regulation	Unknown
TAUS Translation Memory	Unknown
Scale HLE	Unknown
HackerRank Coding	Unknown
RL data for Search	Gemini 3; GPT-5

Private Non-publicly Accessible Datasets by NVIDIA

Dataset	Model(s) used
Simple Minesweeper	Undisclosed
Simple Sudoku	Undisclosed
Multitool Typewriter Hard	Undisclosed
Machine Translation of News Commentary and TAUS Translation Memory	Undisclosed
Machine Translation of STEM -	Qwen2.5-14B-Instruct
Competitive Coding RL data from Nemotron Cascade	Undisclosed
Long context RL	Undisclosed
Single-step SWE RL for patch generation	Undisclosed
OpenHands SWE	Undisclosed

NVIDIA-Sourced Synthetic Datasets (Pre-Training)

Dataset	Modality	Dataset Size	Seed Dataset	Model(s) used for generation
Nemotron-Pretraining-Fact-Seeking	Text	35.0B	FineWiki	Qwen3-30B-A3B-Instruct-2507
Nemotron-Pretraining-Legal	Text	4.3B	CommonPile (caselaw_access_project_filtered); California Code of Regulations; Judicial Ethics Opinions; GLOBALCIT; CUAD; Nemotron Personas; ToSDR Terms of Service Corpus; CodeHima/TOS_Dataset; ContractNLI; CaseHOLD; Code of Federal Regulations; Canadian Case Law (subsets that allow commercial use)	Qwen3-235B-A22B-Thinking-2507
Nemotron-Pretraining-Formal-Logic	Text	128M	Nemotron Personas	Qwen3-235B-A22B-Thinking-2507
Nemotron-Pretraining-Economics	Text	73.4M	-	Qwen3-235B-A22B-Thinking-2507
Nemotron-Pretraining-Multiple-Choice	Text	1.6B	MMLU Auxiliary Train	DeepSeek-V3; Qwen3-235B-A22B
Nemotron-Pretraining-Code-Concepts	Text	7.3B	-	gpt-oss-20b; gpt-oss-120b
Nemotron-Pretraining-Unconditional-Algorithmic	Text	196.5M	-	gpt-oss-120b; Qwen3-235B-A22B
More Synthetic Tasks from DeepSeek-V3 and Qwen3-235B-A22B	Text	1.1B	train splits of acp_bench; ai2_arc; babi; gsm8k; hendrycks_math; IFEval; MedText; mediqa_qa; mlqa; MMLU-Pro; mmlu-pro-plus; MMLU-ProX; nq_open; tinyGSM8k; truthful_qa; truthfulqa-multi; MATH-lighteval; mmlu; awesome-chatgpt-prompts; super_glue	DeepSeek v3; Qwen3-235B-A22B
Synthetic Tasks from DeepSeek-V3 and Qwen3-235B-A22B	Text	6.7B	train splits of Into the Unknown; AI2 ARC (AI2 Reasoning Challenge); BLiMP (Benchmark of Linguistic Minimal Pairs); CommonSenseQA; GLUE; HeadQA; Hendrycks Ethics; Memo Trap; modus-tollens; NeQA; pattern-matching-suppression; mastermind_24_mcq_random; mastermind_24_mcq_close; quote-repetition; redefine-math; Repetitive Algebra; sig-figs; MMLU-Pro; MC-TACO; MedConceptsQA; MMLU_dataset; OpenbooksQA; PIQA (Physical Interaction Question Answering); SocialIQA; SuperGLUE; tinyAI2_arc; tinyMMLU; tinyWinogrande; TruthfulQA; WebQuestions; Winogrande; GPQA; MBPP	DeepSeek v3; Qwen3-235B-A22B
Synthetic Art of Problem Solving from DeepSeek-R1	Text	40B	Art of Problem Solving; American Mathematics Competitions 8; American Mathematics Competitions 10;	DeepSeek-R1
Synthetic Moral Stories and Social Chemistry from Qwen3-235B-A22B-Thinking-2507 and Mixtral-8x22B-v0.1	Text	15.2M	social-chemestry-101; Moral Stories	Qwen3-235B-A22B-Thinking-2507; Mixtral-8x22B-v0.1
Synthetic Moral Stories and Social Chemistry from Mixtral-8x22B-v0.1	Text	327M	social-chemestry-101; Moral Stories	Mixtral-8x22B-v0.1
Synthetic Social Sciences seeded with OpenStax from DeepSeek-V3, Mixtral-8x22B-v0.1, and Qwen2.5-72B	Text	83.6M	OpenStax - CC BY-SA subset	DeepSeek-V3; Mixtral-8x22B-v0.1; Qwen2.5-72B
Synthetic Health Sciences seeded with OpenStax from DeepSeek-V3, Mixtral-8x22B-v0.1, and Qwen2.5-72B	Text	9.7M	OpenStax - CC BY-SA subset	DeepSeek-V3; Mixtral-8x22B-v0.1; Qwen2.5-72B
Synthetic STEM seeded with OpenStax, Open Textbook Library, and GSM8K from DeepSeek-R1, DeepSeek-V3, DeepSeek-V3-0324, and Qwen2.5-72B	Text	175M	OpenStax - CC BY-SA subset; GSM8K; Open Textbook Library - CC BY-SA & GNU subset	DeepSeek-R1, DeepSeek-V3; DeepSeek-V3-0324; Qwen2.5-72B
Nemotron-PrismMath	Text	4.6B	Big-Math-RL-Verified; OpenR1-Math-220k	Qwen2.5-0.5B-instruct, Qwen2.5-72B-Instruct; DeepSeek-R1-Distill-Qwen-32B
Synthetic Question Answering Data from Papers and Permissible Books from Qwen2.5-72B-Instruct	Text	350M	arXiv; National Institutes of Health ExPorter; BioRxiv; PMC Article; USPTO Backgrounds; peS2o; Global Regulation; CORE; PG-19; DOAB CC BY & CC BY-SA subset; NDLTD	Qwen2.5-72B-Instruct
Synthetic Rephrased Math Data from Common Crawl from phi-4	Text	73B	Common Crawl	phi-4
Synthetic Math Data from Common Crawl 4plus	Text	52.3B	Common Crawl	phi-4
Synthetic Math Data from Common Crawl 3	Text	80.9B	Common Crawl	phi-4
Synthetic AGIEval seeded with AQUA-RAT, LogiQA, and AR-LSAT from DeepSeek-V3 and DeepSeek-V3-0324	Text	4.0B	AQUA-RAT; LogiQA; AR-LSAT	DeepSeek-V3; DeepSeek-V3-0324
Synthetic AGIEval seeded with AQUA-RAT, LogiQA, and AR-LSAT from Qwen3-30B-A3B	Text	4.2B	AQUA-RAT; LogiQA; AR-LSAT	Qwen3-30B-A3B
Synthetic Art of Problem Solving from Qwen2.5-32B-Instruct, Qwen2.5-Math-72B, Qwen2.5-Math-7B, and Qwen2.5-72B-Instruct	Text	Undisclosed	Art of Problem Solving; American Mathematics Competitions 8; American Mathematics Competitions 10; GSM8K; PRM800K	Qwen2.5-32B-Instruct; Qwen2.5-Math-72B; Qwen2.5-Math-7B; Qwen2.5-72B-Instruct
Synthetic MMLU Auxiliary Train from DeepSeek-R1	Text	0.5B	MMLU Auxiliary Train	DeepSeek-R1
Synthetic Long Context Continued Post-Training Data from Papers and Permissible Books from Qwen2.5-72B-Instruct	Text	Undisclosed	arXiv; National Institutes of Health ExPorter; BioRxiv; PMC Article; USPTO Backgrounds; peS2o; Global Regulation; CORE; PG-19; DOAB CC BY & CC BY-SA subset; NDLTD	Qwen2.5-72B-Instruct
Synthetic Common Crawl from Qwen3-30B-A3B and Mistral-Nemo-12B-Instruct	Text	415.8B	Common Crawl	Qwen3-30B-A3B; Mistral-NeMo-12B-Instruct
Synthetic Multilingual Data from Common Crawl from Qwen3-30B-A3B	Text	Undisclosed	Common Crawl	Qwen3-30B-A3B
Synthetic Multilingual Data from Wikimedia from Qwen3-30B-A3B	Text	Undisclosed	Wikimedia	Qwen3-30B-A3B
Synthetic Math Data from Wikimedia from Nemotron-4-340B-Instruct	Text	Undisclosed	-	Nemotron-4-340B-Instruct
Synthetic Common Crawl Code from phi-4	Text	427.9B	Common Crawl	phi-4
Synthetic Scientific Coding from Qwen3-235B-A22B	Text	1.2B	Wikimedia	Qwen3-235B-A22B
Tool Calling Data	Text	26.2B		Qwen3-235B-A22B-2507; gpt-oss-120b
Synthetic Essential-Web from QwQ-32B	Text	28.1B	Essential-Web	QwQ-32B
Translated Synthetic Crawl	Text	389.9B	Common Crawl	Qwen3-30B-A3B
Translated Synthetic Wikipedia	Text	7.9B	Wikimedia	Qwen3-30B-A3B
Synthetic Long Context from Qwen3-235B-A22B-Instruct-2507	Text	Undisclosed	CORE; PG-19; DOAB CC BY & CC BY-SA subset; NDLTD	Qwen3-235B-A22B-Instruct-2507
Synthetic Search STEM OPENQ from DeepSeek-R1-0528	Text	Undisclosed	-	DeepSeek-R1-0528
Synthetic MCQ from Qwen2.5-32B-Instruct and DeepSeek-R1-0528	Text	Undisclosed	-	Qwen2.5-32B-Instruct; DeepSeek-R1-0528
Synthetic Offline Search MCQA HLE from DeepSeek-R1-0528	Text	Undisclosed	-	DeepSeek-R1-0528
Synthetic Offline Search MCQA GPQA from Qwen3-235B-A22B and DeepSeek-R1-0528	Text	Undisclosed	-	Qwen3-235B-A22B; DeepSeek-R1-0528
Synthetic Human Preference from QwQ-32B, Qwen3-30B-A3B, Qwen3-235B-A22B, Qwen3-235B-A22B-Instruct-2507, Mistral-Small-3.1-24B-Instruct-2503, Mistral-Small-3.2-24B-Instruct-2506, MiniMax-M1-80k, MiniMax-M1-40k, Kimi-K2-Instruct, DeepSeek-V3-0324, DeepSeek-R1-0528	Text	Undisclosed	-	QwQ-32B; Qwen3-30B-A3B; Qwen3-235B-A22B; Qwen3-235B-A22B-Instruct-2507; Mistral-Small-3.1-24B-Instruct-2503; Mistral-Small-3.2-24B-Instruct-2506; MiniMax-M1-80k; MiniMax-M1-40k; Kimi-K2-Instruct; DeepSeek-V3-0324; DeepSeek-R1-0528
Synthetic Code from Qwen3-32B	Text	Undisclosed	English Common Crawl; English Common Crawl 1.1	Qwen3-32B
Synthetic OpenCodeReasoning from DeepSeek-R1	Text	Undisclosed	OpenCodeReasoning	DeepSeek-R1
Synthetic LIMO from DeepSeek-R1-0528	Text	Undisclosed	LIMO	DeepSeek-R1-0528
Synthetic SCP from DeepSeek-R1-0528	Text	Undisclosed	SCP-116K	DeepSeek-R1-0528
Synthetic Stack Exchange from DeepSeek-R1-0528	Text	Undisclosed	Stack Exchange	DeepSeek-R1-0528
Synthetic Common Crawl from Qwen3-30B-A3B	Text	Undisclosed	Common Crawl	Qwen3-30B-A3B
Synthetic Wikipedia from Qwen3-30B-A3B	Text	Undisclosed	Wikimedia	Qwen3-30B-A3B
Synthetic Essential-Web from Qwen3-30B-A3B and Qwen3-235B-A22B-Thinking-2507	Text	Undisclosed	Essential-Web	Qwen3-30B-A3B; Qwen3-235B-A22B-Thinking-2507
Synthetic Textbook Math from Qwen3-30B-A3B, Qwen3-235B-A22B, phi-4	Text	Undisclosed	Common Crawl; FineMath	Qwen3-30B-A3B; Qwen3-235B-A22B; phi-4
Synthetic Math and Code from DeepSeek-R1 and DeepSeek-R1-0528	Text	Undisclosed	Magicoder-Evol-Instruct-110K; opc-sft-stage2; TACO; OpenCodeReasoning; OpenMathReasoning; NuminaMath CoT	DeepSeek-R1; DeepSeek-R1-0528

NVIDIA-Sourced Synthetic Datasets (Post-Training)

Dataset	Modality	Dataset Size	Seed Dataset	Model(s) used for generation
Synthetic Competitive MATH Proofs from DeepSeek-V4-Pro	Text	Undisclosed	[AMC 8 Problems and Solutions, AMC 10 Problems and Solution, and AIME Problems and Solutions]	[deepseek-ai/DeepSeek-V4-Pro]
Synthetic Hermes Agent Reasoning Traces	Text	Undisclosed	[lambda/hermes-agent-reasoning-traces]	[hermes-agent-generator]
Synthetic Competitive Coding from DeepSeek-V4-Pro	Text	Undisclosed	[NVCompetitiveCodingV1]	[deepseek-ai/DeepSeek-V4-Pro]
Synthetic Competitive Science Reasoning from DeepSeek-V4-Pro	Text	Undisclosed	[AMC 8 Problems and Solutions, AMC 10 Problems and Solution, and AIME Problems and Solutions]; [EssentialAI/essential-web-v1.0]; [cdquestions.com]; [Pile-FreeLaw]; [Vedantu]; [askfilo]; [doubtnut]; [ICHO-IPH0 Dataset]; [LIMO dataset (Less is More for Reasoning)]; [AAPT]; [ChemData 700K]; [oMeBench]; [Flavor Analysis and Recognition Transformer]; [ChemCoTBench]; [Llama Nemotron Dataset]	[deepseek-ai/DeepSeek-V4-Pro]
Synthetic Competitive MATH CoT and TIR from Nemotron 5.5	Text	Undisclosed	[Pile-FreeLaw]; [AMC 8 Problems and Solutions, AMC 10 Problems and Solution, and AIME Problems and Solutions]	[Nemotron 5.5]
Vendor Terminal Bench-like Tasks from Mercor	Text	Undisclosed	[Terminal bench like tasks curated by the vendor]	[Undisclosed - purchased dataset]
Turing Math Data Pack	Text	Undisclosed	[Turing Math Data Pack dataset]	[Undisclosed - purchased dataset]
Synthetic Holdout, Skywork, DAPO, and Turing Math from GPT-5.5	Text	Undisclosed	[DocQA-RL-1.6K]; [DAPO-Math-17k]	[GPT-5.5]
Synthetic Long Context RL from QwenLong L1 and DocQA-RL-1.6K	Text	Undisclosed	[DocQA-RL-1.6K]	Undisclosed
Synthetic Competitive Coding Gym Tasks	Text	Undisclosed	[NVCompetitiveCodingV1.1]	Undisclosed
Synthetic Finance SEC Search Agent from GPT-OSS-120B and Qwen3	Text	Undisclosed	[SEC filings from sec.gov]	[GPT-OSS-120B]; [Qwen3-235B-A22B-Instruct]; [Qwen3-4B-Instruct]
Synthetic Structured Outputs from Qwen3-30B-A3B-Instruct-2507, Qwen3-30B-A3B-Thinking-2507, Qwen3-235B-A22B-Instruct-2507, and Qwen3-235B-A22B-Thinking-2507	Text	Undisclosed	[Nemotron-RL-agent-structured-outputs-v1]	[Qwen3-30B-A3B-Instruct-2507]; [Qwen3-235B-A22B-Instruct-2507]
Synthetic Long Context Equivalence Rule from Qwen3-235B-A22B-Thinking-2507 and DeepSeek-R1	Text	Undisclosed	[Long-context SFT data]	[Qwen/Qwen3-235B-A22B-Thinking-2507]; [Deepseek-ai/DeepSeek-R1]
Synthetic Science RL Data Blend from Qwen2.5-32B	Text	Undisclosed	[doubtnut]; [Pile-FreeLaw]; [Llama Nemotron Dataset]; [askfilo]; [EssentialAI/essential-web-v1.0]; [Vedantu]; [auxiliary_train]; [cdquestions.com]; [AMC 8 Problems and Solutions, AMC 10 Problems and Solution, and AIME Problems and Solutions]; [AAPT]; [ICHO-IPH0 Dataset]; [LIMO dataset (Less is More for Reasoning)]	[Qwen2.5-32B]
Synthetic Abstention Data from Nemotron Super v3	Text	Undisclosed	[Go abstention Dataset]	[nvidia/nvidia/nemotron-3-super-v3]
Synthetic Chemistry Data from Nemotron Super v3	Text	Undisclosed	[ChemData 700K]	[nvidia/nvidia/nemotron-3-super-v3]
Synthetic Structured Outputs from Qwen3-30B-A3B-Instruct-2507, Qwen3-30B-A3B-Thinking-2507, Qwen3-235B-A22B-Instruct-2507, and Qwen3-235B-A22B-Thinking-2507	Text	Undisclosed	[In-house data]	[GPT OSS 120B - Apache 2.0]
Synthetic Tool Call Schema for RL	Text	Undisclosed	[In-house data]	[GPT OSS 120B - Apache 2.0]
Synthetic Freeform Text Formatting from GPT-OSS-120B	Text	Undisclosed	[In-house data]	[GPT OSS 120B - Apache 2.0]
Synthetic Citation Formatting from GPT-OSS-120B	Text	Undisclosed	[In-house data]	[GPT OSS 120B - Apache 2.0]
Droid Harness Pivot Vendor Data	Text	Undisclosed	[Droid Harness Pivot vendor data]	Undisclosed
Synthetic HotpotQA Training Data from Qwen3-235B	Text	Undisclosed	[HotpotQA]	[Qwen3-235B]
Synthetic Natural Language Math Proofs from Nemotron 5.5	Text	Undisclosed	[AMC8, AMC10, and AIME problem sets hosted on Art of Problem Solving]; [Pile-StackExchange]	[Nemotron 5.5]
Synthetic Stack Overflow OpenQ	Text	Undisclosed	[Pile-FreeLaw]	Undisclosed
Chemistry Ether0 Vendor Data	Text	Undisclosed	[Chemistry ether0 vendor data]	Undisclosed
Synthetic Litmus-Bench Chemistry from ChEMBL	Text	Undisclosed	[ChEMBL]; [Nemo Gym RL dataset generated from ChEMBL with RDKit]	Undisclosed
Synthetic ZINC Chemistry from Nemotron Super v3	Text	Undisclosed	[ZINC]	[Nemotron Super v3]
ARC-AGI Gym Environment	Text	Undisclosed	[ARC-AGI-2]	[ARC-AGI-2]
Synthetic Agentic Search Tool-Use from DeepSeek-V3.2	Text	Undisclosed	[Mercor Data]	[DeepSeek-V3.2]
Synthetic Text-To-SQL	Text	Undisclosed	[In-house Text-to-SQL data]	[gpt-oss-120b]
Dialog Memory Vendor Data	Text	Undisclosed	[Patronus external vendor agreement]	Undisclosed
Synthetic Indirect Prompt Injection from Nemotron Super v3 and Qwen3-Next-80B-A3B-Instruct	Text	Undisclosed	[In-house indirect prompt injection data]	[nvidia/nemotron-3-super-v3, qwen/qwen3-next-80b-a3b-instruct.]
Synthetic Malicious Code and Agentic Security	Text	Undisclosed	[In-house malicious-code / agentic-security data]	Undisclosed
Synthetic Single-Step SWE Patch Selection	Text	Undisclosed	[SWE-Gym Dataset]; [SWE Bench Verified Benchmark]	[ground truth and task checks]
Synthetic Natural Language Math Final Answers from Nemotron 5.5	Text	Undisclosed	[AMC8, AMC10, and AIME problem sets hosted on Art of Problem Solving]; [Pile-StackExchange]	[nemotron 5.5]
Synthetic Simple Math Prompts for Token Efficiency	Text	Undisclosed	[In-house simple math prompts]	Undisclosed
Synthetic Abstention Data from Nemotron Super v3	Text	Undisclosed	[CRAG]	[nvidia/nvidia/nemotron-3-super-v3]
Synthetic Agentless SWE	Text	242,536	[SWE-Rebench-V2]; [SWEbench Training Set]; [R2E-Gym/R2E-Gym-Subset]; [SWE-Gym/SWE-Gym]; [SWE-Rebench]	[openai/gpt-oss-120b]
Synthetic Agentic CUDA Traces from GLM-4.7	Text	2,276	[Internal CUDA task data]	[GLM-4.7]
Synthetic Math Proofs from DeepSeek-V3.2-Speciale	Text	820,772	[Nemotron-Math-Proofs-v1]	[SDG: DeepSeek-V3.2-Speciale]; [Filter: proof validation]
Synthetic Multilingual SFT from DeepSeek-V3	Text	1,245,284	[Nano v3 SFT data]	[DeepSeek-V3]
Synthetic Agentic Code from gpt-oss-120b	Text	109,086	[NVAgenticCLIPrompts-v1]; [NVAgenticSkills-v1]; [NVAgenticCLIMultiTurnPrompts-v1]	[openai/gpt-oss-120b]
Synthetic Agentic CLI and Web Skills from gpt-oss-120b	Text	27,418	[NVAgenticCLIPrompts-v1]; [NVAgenticSkills-v1]; [NVAgenticCLIMultiTurnPrompts-v1]; [NVAgenticCLIPrompts-Web-v1]	[openai/gpt-oss-120b]
Synthetic Agentic Coding from gpt-oss-120b	Text	160,531	[NVAgenticCLIPrompts-v1]; [NVAgenticSkills-v1]; [NVAgenticCLIMultiTurnPrompts-v1]; [NVAgenticCLIPrompts-Web-v1]	[openai/gpt-oss-120b]
Synthetic OpenCode Agentic Tasks from gpt-oss-120b	Text	614,773	[NVAgenticCLIPrompts-v1]; [NVAgenticSkills-v1]; [NVAgenticCLIMultiTurnPrompts-v1]; [NVAgenticCLIPrompts-Web-v1]	[openai/gpt-oss-120b]
Synthetic ARC-AGI Ultra Data	Text	192,016	[ARC-AGI-2]; [arc dataset collection]	[ARC-AGI-2]
Synthetic LiveCodeBench TIR from DeepSeek-R1-0528	Text	1,283,398	[Nemotron-X training datasets]	[DeepSeek-R1-0528]
Synthetic Verilog and SystemVerilog Code from DeepSeek-R1-0528 and GPT-OSS-120B	Text	1,233,247	[Verilog/SystemVerilog seed code]	[SDR: DeepSeek R1 0528 and GPT-OSS-120B]; [Filtering: Claude 4 Sonnet]
Synthetic Aider Python Tasks from DeepSeek-R1-0528	Text	236,099	[Exercism (GitHub Python)]	[Deepseek R1 0528]
Synthetic Chat Reasoning-Off Data from GLM-5	Text	646,738	[lmarena-ai/repochat-arena-preference-4k user prompts]	[Multi-turn conversations generated by GLM-5 with best-of-4 selection via Qwen3-Nemotron-235B-A22B-GenRM:]
Synthetic Chat Reasoning-On Data from GLM-5	Text	644,286	[lmarena-ai/repochat-arena-preference-4k user prompts]; [lmarena-ai/arena-expert-5k user prompts]; [lmarena-ai/arena-human-preference-55k user prompts]; [lmarena-ai/arena-human-preference-100k user prompts]; [lmarena-ai/arena-human-preference-140k user prompts]	[Multi-turn conversations generated by GLM-5 with best-of-4 selection via Qwen3-Nemotron-235B-A22B-GenRM:]
Synthetic Multilingual Safety from Riva-Translate-4B-Instruct-v1.1	Text	132,067	[Safety SFT Data: Ultra]	[nvidia/Riva-Translate-4B-Instruct-v1.1]
Synthetic Science Reasoning Effort Medium	Text	502,722	[science-reasoning-effort-medium-v0]	Undisclosed
Synthetic Telecom Tool-Use Trajectories from gpt-oss-120b	Text	12,455	[Existing Tau2 telecom trajectories originally generated with DeepSeek V3.2]	[gpt-oss-120b]
Synthetic Terminal Bench Data from OpenReasoningv2	Text	Undisclosed	[OpenCodeReasoningv2]; [OpenMathReasoning]; [nemo-swe-bench-repos]; [SWE-Rebench]; [SWE-Fixer-110K]	[OpenReasoningv2]
Synthetic Tulu Instruction Following from DeepSeek-R1-0528	Text	105,361	[Nemotron-X training datasets]	[DeepSeek-R1-0528]
Synthetic SWE Unverified	Text	Undisclosed	[NVAgenticCLIPrompts-v1]; [NVAgenticSkills-v1]; [NVAgenticCLIMultiTurnPrompts-v1]; [NVAgenticCLIPrompts-Web-v1]	[gpt-oss-120b]
Synthetic Instruction Following from gpt-oss-120b	Text	151,988	[IFEval]; [IFEvalG]	[gpt-oss-120b]
Synthetic Identity Data from Qwen3-Next-80B-A3B-Instruct and Qwen3-235B-A22B-Instruct-2507	Text	25,992	[Hand-written prompts]	[Qwen3-Next-80B-A3B-Instruct]; [Qwen3-235B-A22B-Instruct-2507]
Synthetic Terminus Ultra Agentic Reasoning Blend	Text	96,881	[ARC-AGI-2]; [OpenCodeReasoningv2]; [OpenMathReasoning]; [SWE-Fixer-110K]; [SWE-Rebench]; [SWE-Smith]	[DeepSeek-V3.2]; [Qwen3-235B-A22B-Thinking-2507]; [Ring-1T]; [Kimi-K2.5]; [GLM-4.7-FP8]; [Qwen3-Next-80B-A3B-Thinking]; [gpt-oss-120b]; [Ministral-3-14B-Reasoning-2512]; [LM-4.5-Air-FP8]
Synthetic STEM from Qwen3-235B-A22B-Thinking-2507	Text	1,174,694	[IChO-IPhO-RL-v2]; [Physics-Big Dataset]	Undisclosed
Translation Data from TAUS	Text	1,618,055	[TAUS proprietary dataset]	Undisclosed
Synthetic Art of Problem Solving and Stack Exchange from gpt-oss-120b, Qwen2.5-32B-Instruct, and Goedel-Prover-V2-32B	Text	860,469	[Nemotron-Math-Proofs-v1]	[Goedel-Prover-V2-32B]
Synthetic Art of Problem Solving and Stack Exchange from gpt-oss-120b, Qwen2.5-32B-Instruct, and Goedel-Prover-V2-32B	Text	1,201,815	[Upstream released math dataset]; [AoPS]; [StackOverflow / StackExchange]	[gpt-oss-120b]
Synthetic Art of Problem Solving and Stack Exchange from gpt-oss-120b, Qwen2.5-32B-Instruct, and Goedel-Prover-V2-32B	Text	1,296,676	[Upstream released math dataset]; [AoPS]; [StackOverflow / StackExchange]	[gpt-oss-120b]
Synthetic Instruction Following for RL	Text	Undisclosed	[WildChat-1M]; [LMSYS-340B-Eval Dataset]; [LMSYS-Chat-1M Prompts]; [IFEval]; [IFEvalG]	[Qwen/Qwen3-235B-A22B-Thinking-2507]; [gpt-oss-120b]; [Qwen3-235B-A22B-Instruct-2507]
Synthetic Instruction Following for RL	Text	Undisclosed	[WildChat-1M]; [LMSYS-340B-Eval Dataset]; [LMSYS-Chat-1M Prompts]; [IFEval]; [IFEvalG]	[Qwen/Qwen3-235B-A22B-Thinking-2507]; [gpt-oss-120b]; [Qwen3-235B-A22B-Instruct-2507]
Synthetic Multilingual Science and Code data from DeepSeek-R1, DeepSeek-R1-0528, Qwen2.5-32B-Instruct, and Qwen3-235B-A22B, translated with Qwen2.5-32B-Instruct and Qwen2.5-14B-Instruct	Text	Undisclosed	[Nano-V3 SFT Data (without tool call)]	[Qwen/Qwen2.5-14B-Instruct]; [Qwen/Qwen3-4B-Thinking-2507]
Synthetic Search Graph Walk	Text	6,977	[Wikidata / Wikipedia KnowledgeBase]	[MiniMaxAI/MiniMax-M2]
Synthetic Agentic Diverse Domains	Text	281,537	[Handwritten prompts (synthetic; no external seed data used)]	[SDG model: deepseek-ai/DeepSeek-V3.2, deepseek-ai/DeepSeek-R1-0528, Qwen/Qwen3-235B-A22B-Thinking-2507, Qwen/Qwen3-32B]; [Filtering model: openai/gpt-oss-120b, Qwen/Qwen3-32B, Qwen/Qwen3-235B-A22B-Instruct-2507]
Synthetic Long Context from Qwen3-235B-A22B-Instruct-2507	Text	65,608	[Long-context SFT seed blend (pre-training blend + nano-v1 post-training data)]	[Qwen/Qwen3-235B-A22B-Thinking-2507 and deepseek-ai/DeepSeek-R1]
Synthetic Agentless SWE	Text	209,976	[SWE-Bench-Train]; [SWE-Fixer-Train]; [SWE-reBench]; [SWE-Smith]	[deepseek-ai/DeepSeek-R1-0528]
Synthetic Nemotron Math SFT from DeepSeek-V3.2-Speciale	Text	1,900,553	[Nemotron-Math-v2 (AOPS and StackExchange-math problems)]	[DeepSeek-V3.2-Speciale]
Synthetic Nemotron Math TIR from DeepSeek-V3.2	Text	1,789,258	[Nemotron-Math-v2 (AOPS and StackExchange-math problems)]	[DeepSeek-V3.2]
Synthetic SWE Unverified	Text	27,911	[NVAgenticCLIPrompts-v1]; [NVAgenticSkills-v1]	[gpt-oss-120b]
Synthetic SWE Unverified	Text	28,116	[NVAgenticCLIPrompts-v1]; [NVAgenticSkills-v1]	[Qwen3-Coder-480B-A35B-Instruct]
Synthetic NemoCascade OCR Distillation from gpt-oss-120b	Text	682,864	[Nemotron-X training datasets]	[gpt-oss-120b]
Synthetic SWE Unverified	Text	26,865	[NVAgenticCLIPrompts-v1]; [NVAgenticSkills-v1]	[gpt-oss-120b]; [Qwen/Qwen3-Coder-480B-A35B-Instruct]; [GLM-4.7-Flash]
Synthetic CUDA 100k	Text	93,086	[KernelBook]; [HuggingFace Transformers]; [FlashInfer]	[gpt-oss-120b]; [DeepSeek-R1-0528]
Synthetic Science MCQ and QA Diversity from GPT-OSS and Kimi-K2	Text	30,358	[doubtnut]; [Pile-FreeLaw]; [Llama Nemotron Dataset]; [askfilo]; [EssentialAI/essential-web-v1.0]; [Vedantu]; [auxiliary_train]; [cdquestions.com]; [AMC 8 Problems and Solutions, AMC 10 Problems and Solution, and AIME Problems and Solutions]; [AAPT]; [ICHO-IPH0 Dataset]; [LIMO dataset (Less is More for Reasoning)]	[GPT-OSS]; [Kimi-K2]
Synthetic Science HLE with Python from GPT-OSS and Kimi-K2	Text	85,184	[doubtnut]; [Pile-FreeLaw]; [Llama Nemotron Dataset]; [askfilo]; [EssentialAI/essential-web-v1.0]; [Vedantu]; [auxiliary_train]; [cdquestions.com]; [AMC 8 Problems and Solutions, AMC 10 Problems and Solution, and AIME Problems and Solutions]; [AAPT]; [ICHO-IPH0 Dataset]; [LIMO dataset (Less is More for Reasoning)]	[GPT-OSS]; [Kimi-K2]
Synthetic Science Search and Python from GPT-OSS and Kimi-K2	Text	6,179	[doubtnut]; [Pile-FreeLaw]; [Llama Nemotron Dataset]; [askfilo]; [EssentialAI/essential-web-v1.0]; [Vedantu]; [auxiliary_train]; [cdquestions.com]; [AMC 8 Problems and Solutions, AMC 10 Problems and Solution, and AIME Problems and Solutions]; [AAPT]; [ICHO-IPH0 Dataset]; [LIMO dataset (Less is More for Reasoning)]	[GPT-OSS]; [Kimi-K2]
Synthetic Science Search from GPT-OSS and Kimi-K2	Text	32,554	[doubtnut]; [Pile-FreeLaw]; [Llama Nemotron Dataset]; [askfilo]; [EssentialAI/essential-web-v1.0]; [Vedantu]; [auxiliary_train]; [cdquestions.com]; [AMC 8 Problems and Solutions, AMC 10 Problems and Solution, and AIME Problems and Solutions]; [AAPT]; [ICHO-IPH0 Dataset]; [LIMO dataset (Less is More for Reasoning)]	[GPT-OSS]; [Kimi-K2]
Synthetic Finance Reasoning from GPT-OSS-120B and Qwen3-235B-A22B-Instruct-2507	Text	326,700	[_SEC filings]	[GPT-OSS-120B, Qwen3-235B-A22B-Instruct-2507]
Synthetic Science Diversity MCQ from GPT-OSS and Kimi-K2	Text	532,942	[doubtnut]; [Pile-FreeLaw]; [Llama Nemotron Dataset]; [askfilo]; [EssentialAI/essential-web-v1.0]; [Vedantu]; [auxiliary_train]; [cdquestions.com]; [AMC 8 Problems and Solutions, AMC 10 Problems and Solution, and AIME Problems and Solutions]; [AAPT]; [ICHO-IPH0 Dataset]; [LIMO dataset (Less is More for Reasoning)]	[GPT-OSS]; [Kimi-K2]
Synthetic Science Diversity OpenQ from GPT-OSS and Kimi-K2	Text	131,045	[doubtnut]; [Pile-FreeLaw]; [Llama Nemotron Dataset]; [askfilo]; [EssentialAI/essential-web-v1.0]; [Vedantu]; [auxiliary_train]; [cdquestions.com]; [AMC 8 Problems and Solutions, AMC 10 Problems and Solution, and AIME Problems and Solutions]; [AAPT]; [ICHO-IPH0 Dataset]; [LIMO dataset (Less is More for Reasoning)]	[GPT-OSS]; [Kimi-K2]
Synthetic Science Reasoning No-Tool from GPT-OSS and Kimi-K2	Text	2,085,600	[doubtnut]; [Pile-FreeLaw]; [Llama Nemotron Dataset]; [askfilo]; [EssentialAI/essential-web-v1.0]; [Vedantu]; [auxiliary_train]; [cdquestions.com]; [AMC 8 Problems and Solutions, AMC 10 Problems and Solution, and AIME Problems and Solutions]; [AAPT]; [ICHO-IPH0 Dataset]; [LIMO dataset (Less is More for Reasoning)]	[GPT-OSS]; [Kimi-K2]
Synthetic Long Context from Qwen3-235B-A22B-Instruct-2507	Text	62,333	[Long-context SFT data: lc_nothink 256k]	[Qwen/Qwen3-235B-A22B-Thinking-2507 and deepseek-ai/DeepSeek-R1]
Synthetic Long Context from Qwen3-235B-A22B-Instruct-2507	Text	49,698	[Long-context SFT data: MRCR 200k]	[Qwen/Qwen3-235B-A22B-Thinking-2507 and deepseek-ai/DeepSeek-R1]
Synthetic Text-To-SQL	Text	96,564	[Undisclosed - no seed data listed]	[gpt-oss-120b]
Synthetic Long Context from Qwen3-235B-A22B-Instruct-2507	Text	397,538	[Long-context SFT data: RULER 256k]	[Qwen/Qwen3-235B-A22B-Thinking-2507 and deepseek-ai/DeepSeek-R1]
Synthetic SWE Unverified	Text	27,960	[NVAgenticCLIPrompts-v1]; [NVAgenticSkills-v1]	[gpt-oss-120b]; [Qwen/Qwen3-Coder-480B-A35B-Instruct]; [GLM-4.7-Flash]
Synthetic SWE Unverified	Text	24,632	[NVAgenticCLIPrompts-v1]; [NVAgenticSkills-v1]	[gpt-oss-120b]; [Qwen/Qwen3-Coder-480B-A35B-Instruct]; [GLM-4.7-Flash]
Synthetic Long Context from Qwen3-235B-A22B-Instruct-2507	Text	49,902	[Long-context SFT data]	[Qwen/Qwen3-235B-A22B-Thinking-2507 and deepseek-ai/DeepSeek-R1]
Synthetic Tool Call Schema for RL	Text	469,983	[UltraTool]; [ToolEyes]; [AutoTools]; [API-Bank]; [Nemotron-Personas-USA]; [Salesforce xLAM function-calling]; [Glaive function-calling-v2]; [Agent-Ark/Toucan-1.5M]	[DeepSeek-V3.2]; [GLM-4.6]; [gpt-oss-120b]; [Kimi-K2-Instruct]
Synthetic Tool Call Schema for RL	Text	707,967	[UltraTool]; [ToolEyes]; [AutoTools]; [API-Bank]; [Nemotron-Personas-USA]; [Salesforce xLAM function-calling]; [Glaive function-calling-v2]; [Agent-Ark/Toucan-1.5M]	[DeepSeek-V3.2]; [GLM-4.6]; [gpt-oss-120b]; [Kimi-K2-Instruct]
Synthetic Long Context from Qwen3-235B-A22B-Instruct-2507	Text	52,630	[AALCR seed blend: SEC Filings]; [CC]; [Wikipedia]; [FinePDFs]; [ArXiv]; [Pile-NIH ExPorter]; [BioRxiv]; [PMC Article]; [USPTO Backgrounds]; [peS20]; [Global Regulations]; [CORE]; [Gutenberg (PG-19)]; [DOAB CC-BY]; [NDLTD]; [Amps]; [StackExchange]; [MathPile]; [Numinas]	[Qwen3-30B-A3B]
Synthetic Safety from gemma-3-4b-it, Nemotron-Nano-9B-v2, and gpt-oss-120b	Text	44,091	[Safety SFT Data]	[google/gemma-3-4b-it]; [Nemotron-Nano-9B-v2]; [gpt-oss-120b]

Language Distribution in Post-Training

For our post-training recipe, we focused on the following languages in addition to English: French, Spanish, Italian, German, Japanese, Korean, Hindi, Korean, Brazilian Portuguese, and Chinese

Those languages were represented in the form of multilingual reasoning and translation tasks.

The following table depicts our sample distribution.

Language	Size
English	8.6M
Italian	138k
German	138k
Spanish	138k
French	138k
Japanese	138k
Chinese	138k
Hindi	138k
Korean	138k
Brazilian Portuguese	138k

Evaluation Datasets:

Data Collection Method by dataset

Hybrid: Automated, Human, Synthetic

Labeling Method by dataset

Hybrid: Automated, Human, Synthetic

Properties: This corpus comprises a mix of high-quality standard benchmarks and test suites for modern agentic AI as outlined in the benchmark section of the model card.

Testing Datasets:

Data Collection Method by dataset

Hybrid: Automated, Human, Synthetic

Labeling Method by dataset

Hybrid: Automated, Human, Synthetic

Properties: This corpus comprises a mix of high-quality standard benchmarks and test suites for modern agentic AI as outlined in the benchmark section of the model card.

Inference

Acceleration Engine: PyTorch
Test Hardware:
- NVIDIA Hopper
  - H100
  - H200
- NVIDIA Grace Blackwell
  - GB200
  - GB300
- NVIDIA Blackwell
  - B200
  - B300

Ethical Considerations

NVIDIA believes Trustworthy AI is a shared responsibility and we have established policies and practices to enable development for a wide array of AI applications. When downloaded or used in accordance with our terms of service, developers should work with their internal model team to ensure this model meets requirements for the relevant industry and use case and addresses unforeseen product misuse.

We advise against circumvention of any provided safety guardrails contained in the Model without a substantially similar guardrail appropriate for your use case. For more details: Safety and Explainability Subcards.

For more detailed information on ethical considerations for this model, please see the Model Card++ Bias, and Privacy Subcards.

Please report model quality, risk, security vulnerabilities or NVIDIA AI Concerns here.

Citation

@misc{nvidia_nemotron_3_ultra_2026,
  title  = {Nemotron 3 Ultra: Open, Efficient Mixture-of-Experts Hybrid Mamba-Transformer Model for Agentic Reasoning},
  author = {{NVIDIA}},
  year   = {2026},
  url    = {https://research.nvidia.com/labs/nemotron/files/NVIDIA-Nemotron-3-Ultra-Technical-Report.pdf},
  note   = {White Paper}
}