🚀 New models by Bria.ai, generate and edit images at scale 🚀
nvidia/
NVIDIA-Nemotron-Nano-9B-v2
$0.04
in
$0.16
out
NVIDIA-Nemotron-Nano-9B-v2 is a large language model (LLM) trained from scratch by NVIDIA, and designed as a unified model for both reasoning and non-reasoning tasks. 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 controlled via a system prompt. If the user prefers the model to provide its final answer without intermediate reasoning traces, it can be configured to do so.
NVIDIA-Nemotron-Nano-9B-v2
Ask me anything
Settings
NVIDIA-Nemotron-Nano-9B-v2
Model Developer: NVIDIA Corporation
Model Dates:
June 2025 - August 2025
Data Freshness:
September 2024
The pretraining data has a cutoff date of September 2024.
Model Overview
NVIDIA-Nemotron-Nano-9B-v2 is a large language model (LLM) trained from scratch by NVIDIA, and designed as a unified model for both reasoning and non-reasoning tasks. 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 controlled via a system prompt. If the user prefers the model to provide its final answer without intermediate reasoning traces, it can be configured to do so, albeit with a slight decrease in accuracy for harder prompts that require reasoning. Conversely, allowing the model to generate reasoning traces first generally results in higher-quality final solutions to queries and tasks.
The model uses a hybrid architecture consisting primarily of Mamba-2 and MLP layers combined with just four Attention layers. For the architecture, please refer to the Nemotron-H tech report. The model was trained using Megatron-LM and NeMo-RL.
The supported languages include: English, German, Spanish, French, Italian, and Japanese. Improved using Qwen.
This model is ready for commercial use.
License/Terms of Use
GOVERNING TERMS: This trial service is governed by the NVIDIA API Trial Terms of Service. Use of this model is governed by the NVIDIA Open Model License Agreement.
Evaluation Results
Benchmark Results (Reasoning On)
We evaluated our model in Reasoning-On mode across all benchmarks, except RULER, which is evaluated in Reasoning-Off mode.
| Benchmark | Qwen3-8B | NVIDIA-Nemotron-Nano-9B-v2 |
|---|---|---|
| AIME25 | 69.3% | 72.1% |
| MATH500 | 96.3% | 97.8% |
| GPQA | 59.6% | 64.0% |
| LCB | 59.5% | 71.1% |
| BFCL v3 | 66.3% | 66.9% |
| IFEval (Instruction Strict) | 89.4% | 90.3% |
| HLE | 4.4% | 6.5% |
| RULER (128K) | 74.1% | 78.9% |
All evaluations were done using NeMo-Skills. We published a tutorial with all details necessary to reproduce our evaluation results.
Model Architecture
- Architecture Type: Mamba2-Transformer Hybrid
- Network Architecture: Nemotron-Hybrid
Deployment Geography: Global
Use Case
NVIDIA-Nemotron-Nano-9B-v2 is a general purpose reasoning and chat model intended to be used in English and coding languages. Other non-English languages (German, French, Italian, Spanish and Japanese) are also supported. Developers designing AI Agent systems, chatbots, RAG systems, and other AI-powered applications. Also suitable for typical instruction-following tasks.
Release Date: 08/18/2025
- Huggingface 08/18/2025 via https://huggingface.co/nvidia/NVIDIA-Nemotron-Nano-9B-v2
- API Catalog 08/18/2025 via https://build.nvidia.com/nvidia/nvidia-nemotron-nano-9b-v2
References
Input
- Input Type(s): Text
- Input Format(s): String
- Input Parameters: One-Dimensional (1D): Sequences
- Other Properties Related to Input: Context length up to 128K. Supported languages include German, Spanish, French, Italian, Korean, Portuguese, Russian, Japanese, Chinese and English.
Output
- Output Type(s): Text
- Output Format: String
- Output Parameters: One-Dimensional (1D): Sequences up to 128K
Our 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 25.07.nemotron-nano-v2
- Supported Hardware Microarchitecture Compatibility: NVIDIA A10G, NVIDIA H100-80GB, NVIDIA A100
- Operating System(s): Linux
Case 1: /think or no reasoning signal is provided in the system prompt, reasoning will be set to True
messages = [
{"role": "system", "content": "/think"},
{"role": "user", "content": "Write a haiku about GPUs"},
]
Case 2: /no_think is provided, reasoning will be set to False
messages = [
{"role": "system", "content": "/no_think"},
{"role": "user", "content": "Write a haiku about GPUs"},
]
Note: /think or /no_think keywords can also be provided in “user” messages for turn-level reasoning control.
The rest of the inference snippet remains the same
We recommend setting temperature to 0.6, top_p to 0.95 for reasoning True and greedy search for reasoning False, and increase max_new_tokens to 1024 or higher for reasoning True.
Prompt Format
We follow the jinja chat template provided below. This template conditionally adds **\<think>**\n to the start of the Assistant response if /think is found in either the system prompt or any user message. If no reasoning signal is added, the model defaults to reasoning "on" mode. The chat template adds <think>**\</think>** to the start of the Assistant response if /no_think is found in the system prompt. Thus enforcing reasoning on/off behavior.
{%- set ns = namespace(enable_thinking = true) %}
{%- for message in messages -%}
{%- set content = message['content'] -%}
{%- if message['role'] == 'user' or message['role'] == 'system' -%}
{%- if '/think' in content -%}
{%- set ns.enable_thinking = true -%}
{%- elif '/no_think' in content -%}
{%- set ns.enable_thinking = false -%}
{%- endif -%}
{%- endif -%}
{%- endfor -%}
{%- if messages[0]['role'] != 'system' -%}
{%- set ns.non_tool_system_content = '' -%}
{{- '<SPECIAL_10>System\n' -}}
{%- else -%}
{%- set ns.non_tool_system_content = messages[0]['content']
.replace('/think', '')
.replace('/no_think', '')
.strip()
-%}
{{- '<SPECIAL_10>System\n' + ns.non_tool_system_content }}
{%- endif -%}
{%- if tools -%}
{%- if ns.non_tool_system_content is defined and ns.non_tool_system_content != '' -%}
{{- '\n\n' -}}
{%- endif -%}
{{- 'You can use the following tools to assist the user if required:' -}}
{{- '\n<AVAILABLE_TOOLS>[' -}}
{%- for tool in tools -%}
{{- (tool.function if tool.function is defined else tool) | tojson -}}
{{- ', ' if not loop.last else '' -}}
{%- endfor -%}
{{- ']</AVAILABLE_TOOLS>\n\n' -}}
{{- 'If you decide to call any tool(s), use the following format:\n' -}}
{{- '<TOOLCALL>[{{"name": "tool_name1", "arguments": "tool_args1"}}, ' -}}
{{- '{{"name": "tool_name2", "arguments": "tool_args2"}}]</TOOLCALL>\n\n' -}}
{{- 'The user will execute tool-calls and return responses from tool(s) in this format:\n' -}}
{{- '<TOOL_RESPONSE>[{{"tool_response1"}}, {{"tool_response2"}}]</TOOL_RESPONSE>\n\n' -}}
{{- 'Based on the tool responses, you can call additional tools if needed, correct tool calls if any errors are found, or just respond to the user.' -}}
{%- endif -%}
{{- '\n' -}}
{%- set messages = messages[1:] if messages[0]['role'] == 'system' else messages -%}
{%- if messages[-1]['role'] == 'assistant' -%}
{%- set ns.last_turn_assistant_content = messages[-1]['content'].strip() -%}
{%- set messages = messages[:-1] -%}
{%- endif -%}
{%- for message in messages -%}
{%- set content = message['content'] -%}
{%- if message['role'] == 'user' -%}
{{- '<SPECIAL_11>User\n' + content.replace('/think', '').replace('/no_think', '').strip() + '\n' }}
{%- elif message['role'] == 'tool' -%}
{%- if loop.first or (messages[loop.index0 - 1].role != 'tool') -%}
{{- '<SPECIAL_11>User\n' + '<TOOL_RESPONSE>[' }}
{%- endif -%}
{{- message['content'] -}}
{{- ', ' if not loop.last and (messages[loop.index0 + 1].role == 'tool') else '' -}}
{%- if loop.last or (messages[loop.index0 + 1].role != 'tool') -%}
{{- ']</TOOL_RESPONSE>\n' -}}
{%- endif -%}
{%- elif message['role'] == 'assistant' -%}
{%- if '</think>' in content -%}
{%- set content = content.split('</think>')[1].strip() %}
{%- endif -%}
{{- '<SPECIAL_11>Assistant\n' + content.strip() }}
{%- if message.tool_calls -%}
{%- if content.strip() != '' -%}
{{- '\n\n' -}}
{%- endif -%}
{{- '<TOOLCALL>[' -}}
{%- for call in message.tool_calls -%}
{%- set fn = call.function if call.function is defined else call -%}
{{- '{"name": "' + fn.name + '", "arguments": ' -}}
{%- if fn.arguments is string -%}
{{- fn.arguments -}}
{%- else -%}
{{- fn.arguments | tojson -}}
{%- endif -%}
{{- '}' + (', ' if not loop.last else '') -}}
{%- endfor -%}
{{- ']</TOOLCALL>' -}}
{%- endif -%}
{{- '\n<SPECIAL_12>\n' -}}
{%- endif -%}
{%- endfor -%}
{%- if add_generation_prompt -%}
{{- '<SPECIAL_11>Assistant\n' -}}
{%- if ns.enable_thinking is defined and ns.enable_thinking is false -%}
{{- '<think></think>' -}}
{%- else -%}
{{- '<think>\n' -}}
{%- endif -%}
{%- if ns.last_turn_assistant_content is defined and ns.last_turn_assistant_content != '' -%}
{{- ns.last_turn_assistant_content -}}
{%- endif -%}
{%- else -%}
{%- if ns.last_turn_assistant_content is defined and ns.last_turn_assistant_content != '' -%}
{{- '<SPECIAL_11>Assistant\n' -}}
{%- if ns.enable_thinking is defined and ns.enable_thinking is false -%}
{{- '<think></think>' -}}
{%- else -%}
{{- '<think>\n' -}}
{%- endif -%}
{{- ns.last_turn_assistant_content -}}
{%- if continue_final_message is defined -%}
{%- if continue_final_message is false -%}
{{- '\n<SPECIAL_12>\n' -}}
{%- endif -%}
{%- else -%}
{{- '\n<SPECIAL_12>\n' -}}
{%- endif -%}
{%- endif -%}
{%- endif -%}
Training, Testing, and Evaluation Datasets
Training datasets
- Data Modality: Text
- Text Training Data Size: More than 10 Trillion Tokens
- Train/Test/Valid Split: We used 100% of the corpus for pre-training and relied on external benchmarks for testing.
- Data Collection Method by dataset: Hybrid: Automated, Human, Synthetic
- Labeling Method by dataset: Hybrid: Automated, Human, Synthetic
Properties: The post-training corpus for NVIDIA-Nemotron-Nano-9B-v2 consists of English and multilingual text (German, Spanish, French, Italian, Korean, Portuguese, Russian, Japanese, Chinese and English). Our sources cover a variety of document types such as: webpages, dialogue, articles, and other written materials. The corpus spans domains including code, legal, math, science, finance, and more. We also include a small portion of question-answering, and alignment style data to improve model accuracies. For several of the domains listed above we used synthetic data, specifically reasoning traces, from DeepSeek R1/R1-0528, Qwen3-235B-A22B, Nemotron 4 340B, Qwen2.5-32B-Instruct-AWQ, Qwen2.5-14B-Instruct, Qwen 2.5 72B.
The pre-training corpus for NVIDIA-Nemotron-Nano-9B-v2 consists of high-quality curated and synthetically-generated data. It is trained in the English language, as well as 15 multilingual 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 twenty trillion tokens.
Alongside the model, we release our final pretraining 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 Nano 2: An Accurate and Efficient Hybrid Mamba-Transformer Reasoning Model .
Public Datasets
Private Non-publicly Accessible Datasets of Third Parties
| Dataset |
|---|
| Global Regulation |
| Workbench |
Online Dataset Sources
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 (Tokens) | Collection Period |
|---|---|---|---|
| English Common Crawl | Text | 3.360T | 4/8/2025 |
| Multilingual Common Crawl | Text | 812.7B | 5/1/2025 |
| GitHub Crawl | Text | 747.4B | 4/29/2025 |
NVIDIA-Sourced Synthetic Datasets
| Dataset | Modality | Dataset Size (Tokens) | Seed Dataset | Model(s) used for generation |
|---|---|---|---|---|
| Synthetic Art of Problem Solving from DeepSeek-R1 | Text | 25.5B | Art of Problem Solving; American Mathematics Competitions 8; American Mathematics Competitions 10; | DeepSeek-R1 |
| 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 FineMath-4+ Reprocessed from DeepSeek-V3 | Text | 9.2B | Common Crawl | DeepSeek-V3 |
| Synthetic FineMath-3+ Reprocessed from phi-4 | Text | 27.6B | Common Crawl | phi-4 |
| Synthetic Union-3+ Reprocessed from phi-4 | Text | 93.1B | Common Crawl | phi-4 |
| Refreshed Nemotron-MIND from phi-4 | Text | 73B | Common Crawl | phi-4 |
| Synthetic Union-4+ Reprocessed from phi-4 | Text | 14.12B | Common Crawl | phi-4 |
| Synthetic Union-3+ minus 4+ Reprocessed from phi-4 | Text | 78.95B | Common Crawl | phi-4 |
| Synthetic Union-3 Refreshed from phi-4 | Text | 80.94B | Common Crawl | phi-4 |
| Synthetic Union-4+ Refreshed from phi-4 | Text | 52.32B | 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 | 83.1B | 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 | 5.4B | 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 | 1.949T | Common Crawl | Qwen3-30B-A3B; Mistral-NeMo-12B-Instruct |
| Synthetic Multilingual Data from Common Crawl from Qwen3-30B-A3B | Text | 997.3B | Common Crawl | Qwen3-30B-A3B |
| Synthetic Multilingual Data from Wikimedia from Qwen3-30B-A3B | Text | 55.1B | Wikimedia | Qwen3-30B-A3B |
| Synthetic OpenMathReasoning from DeepSeek-R1-0528 | Text | 1.5M | OpenMathReasoning | DeepSeek-R1-0528 |
| Synthetic OpenCodeReasoning from DeepSeek-R1-0528 | Text | 1.1M | OpenCodeReasoning | DeepSeek-R1-0528 |
| Synthetic Science Data from DeepSeek-R1-0528 | Text | 1.5M | - | DeepSeek-R1-0528 |
| Synthetic Humanity's Last Exam from DeepSeek-R1-0528 | Text | 460K | Humanity's Last Exam | DeepSeek-R1-0528 |
| Synthetic ToolBench from Qwen3-235B-A22B | Text | 400K | ToolBench | Qwen3-235B-A22B |
| Synthetic Nemotron Content Safety Dataset V2, eval-safety, Gretel Synthetic Safety Alignment, and RedTeam_2K from DeepSeek-R1-0528 | Text | 52K | Nemotron Content Safety Dataset V2; eval-safety; Gretel Synthetic Safety Alignment; RedTeam_2K | DeepSeek-R1-0528 |
| Synthetic HelpSteer from Qwen3-235B-A22B | Text | 120K | HelpSteer3; HelpSteer2 | Qwen3-235B-A22B |
| Synthetic Alignment data from Mixtral-8x22B-Instruct-v0.1, Mixtral-8x7B-Instruct-v0.1, and Nemotron-4 Family | Text | 400K | HelpSteer2; C4; LMSYS-Chat-1M; ShareGPT52K; tigerbot-kaggle-leetcodesolutions-en-2k; GSM8K; PRM800K; lm_identity (NVIDIA internal); FinQA; WikiTableQuestions; Riddles; ChatQA nvolve-multiturn (NVIDIA internal); glaive-function-calling-v2; SciBench; OpenBookQA; Advanced Reasoning Benchmark; Public Software Heritage S3; Khan Academy Math Keywords | Nemotron-4-15B-Base (NVIDIA internal); Nemotron-4-15B-Instruct (NVIDIA internal); Nemotron-4-340B-Base; Nemotron-4-340B-Instruct; Nemotron-4-340B-Reward; Mixtral-8x7B-Instruct-v0.1; Mixtral-8x22B-Instruct-v0.1 |
| Synthetic LMSYS-Chat-1M from Qwen3-235B-A22B | Text | 1M | LMSYS-Chat-1M | Qwen3-235B-A22B |
| Synthetic Multilingual Reasoning data from DeepSeek-R1-0528, Qwen2.5-32B-Instruct-AWQ, and Qwen2.5-14B-Instruct | Text | 25M | OpenMathReasoning; OpenCodeReasoning | DeepSeek-R1-0528; Qwen2.5-32B-Instruct-AWQ (translation); Qwen2.5-14B-Instruct (translation); |
| Synthetic Multilingual Reasoning data from Qwen3-235B-A22B and Gemma 3 Post-Trained models | Text | 5M | WildChat | Qwen3-235B-A22B; Gemma 3 PT 12B; Gemma 3 PT 27B |
Evaluation Dataset:
- Data Collection Method by dataset: Hybrid: Human, Synthetic
- Labeling Method by dataset: Hybrid: Automated, Human, Synthetic
Inference
-
Engines: HF, vLLM, TRT-LLM
-
Test Hardware NVIDIA A10G 24GB, H100 80GB
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 Trustworthy AI 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.
For more detailed information on ethical considerations for this model, please see the Model Card++ Bias, Explainability, Safety & Security, and Privacy Subcards.
Please report security vulnerabilities or NVIDIA AI Concerns here.
Citation
@misc{nvidia2025nvidianemotronnano2,
title={NVIDIA Nemotron Nano 2: An Accurate and Efficient Hybrid Mamba-Transformer Reasoning Model},
author={NVIDIA},
year={2025},
eprint={2508.14444},
archivePrefix={arXiv},
primaryClass={cs.CL},
url={https://arxiv.org/abs/2508.14444},
}
© 2025 Deep Infra. All rights reserved.