README
AceReason-Math Dataset
Overview
AceReason-Math is a high quality, verfiable, challenging and diverse math dataset for training math reasoning model using reinforcement leraning. This dataset contains
- 49K math problems and answer sourced from NuminaMath and DeepScaler-Preview
- applying filtering rules to exclude unsuitable data (e.g., multiple sub-questions, multiple-choice, true/false, long and complex answers, proof, figure)
- this dataset was used to train AceReason-Nemotron models, which achieve strong results on math benchmark such as AIME24 and AIME25.
| Model | AIME 2024 (avg@64) | AIME 2025 (avg@64) |
|---|---|---|
| QwQ-32B | 79.5 | 65.8 |
| DeepSeek-R1-671B | 79.8 | 70.0 |
| Llama-Nemotron-Ultra-253B | 80.8 | 72.5 |
| o3-mini (medium) | 79.6 | 76.7 |
| Light-R1-14B | 74 | 60.2 |
| OpenMath-Nemotron-14B | 76.3 | 63.0 |
| Llama-Nemotron-Super-49B-v1 | 67.5 | 60.0 |
| DeepSeek-R1-Distilled-Qwen-14B | 69.7 | 50.2 |
| DeepSeek-R1-Distilled-Qwen-32B | 72.6 | 54.9 |
| AceReason-Nemotron-7B 🤗 | 69.0 | 53.6 |
| AceReason-Nemotron-14B 🤗 | 78.6 | 67.4 |
Correspondence to
Yang Chen (yachen@nvidia.com), Zhuolin Yang (zhuoliny@nvidia.com), Zihan Liu (zihanl@nvidia.com), Chankyu Lee (chankyul@nvidia.com), Wei Ping (wping@nvidia.com)
License/Terms of Use:
Governing Terms: This dataset is licensed under the Creative Commons Attribution 4.0 International License (CC BY 4.0) available at https://creativecommons.org/licenses/by/4.0/legalcode.
Data Developer:
NVIDIA
Intended Usage:
The AceReason-Math Dataset is intended to be used by the community to deploy reinforcement learning with LLMs. The data may be used to train and evaluate.
Release Date:
6/2/2025
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.
Please report security vulnerabilities or NVIDIA AI Concerns here.
Citation
@article{chen2025acereason,
title={AceReason-Nemotron: Advancing Math and Code Reasoning through Reinforcement Learning},
author={Chen, Yang and Yang, Zhuolin and Liu, Zihan and Lee, Chankyu and Xu, Peng and Shoeybi, Mohammad and Catanzaro, Bryan and Ping, Wei},
journal={arXiv preprint arXiv:2505.16400},
year={2025}
}
Similar repos
README
AceReason-Math Dataset
Overview
AceReason-Math is a high quality, verfiable, challenging and diverse math dataset for training math reasoning model using reinforcement leraning. This dataset contains
- 49K math problems and answer sourced from NuminaMath and DeepScaler-Preview
- applying filtering rules to exclude unsuitable data (e.g., multiple sub-questions, multiple-choice, true/false, long and complex answers, proof, figure)
- this dataset was used to train AceReason-Nemotron models, which achieve strong results on math benchmark such as AIME24 and AIME25.
| Model | AIME 2024 (avg@64) | AIME 2025 (avg@64) |
|---|---|---|
| QwQ-32B | 79.5 | 65.8 |
| DeepSeek-R1-671B | 79.8 | 70.0 |
| Llama-Nemotron-Ultra-253B | 80.8 | 72.5 |
| o3-mini (medium) | 79.6 | 76.7 |
| Light-R1-14B | 74 | 60.2 |
| OpenMath-Nemotron-14B | 76.3 | 63.0 |
| Llama-Nemotron-Super-49B-v1 | 67.5 | 60.0 |
| DeepSeek-R1-Distilled-Qwen-14B | 69.7 | 50.2 |
| DeepSeek-R1-Distilled-Qwen-32B | 72.6 | 54.9 |
| AceReason-Nemotron-7B 🤗 | 69.0 | 53.6 |
| AceReason-Nemotron-14B 🤗 | 78.6 | 67.4 |
Correspondence to
Yang Chen (yachen@nvidia.com), Zhuolin Yang (zhuoliny@nvidia.com), Zihan Liu (zihanl@nvidia.com), Chankyu Lee (chankyul@nvidia.com), Wei Ping (wping@nvidia.com)
License/Terms of Use:
Governing Terms: This dataset is licensed under the Creative Commons Attribution 4.0 International License (CC BY 4.0) available at https://creativecommons.org/licenses/by/4.0/legalcode.
Data Developer:
NVIDIA
Intended Usage:
The AceReason-Math Dataset is intended to be used by the community to deploy reinforcement learning with LLMs. The data may be used to train and evaluate.
Release Date:
6/2/2025
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.
Please report security vulnerabilities or NVIDIA AI Concerns here.
Citation
@article{chen2025acereason,
title={AceReason-Nemotron: Advancing Math and Code Reasoning through Reinforcement Learning},
author={Chen, Yang and Yang, Zhuolin and Liu, Zihan and Lee, Chankyu and Xu, Peng and Shoeybi, Mohammad and Catanzaro, Bryan and Ping, Wei},
journal={arXiv preprint arXiv:2505.16400},
year={2025}
}