Total runs:
65.2K
Run Growth:
12.7K
Growth Rate:
19.52%
Sponsored by
ThumbnailCreator.com
Skywork / Skywork-OR1-32B-Preview
Model's Last Updated: 2025年5月29日
Introduction of Skywork-OR1-32B-Preview
Model Details of Skywork-OR1-32B-Preview
🔥 News
-
April 13, 2025
: We release the
Skywork-OR1(Open Reasoner 1) series of models, includingSkywork-OR1-Math-7B,Skywork-OR1-32B-Preview, andSkywork-OR1-7B-Preview. We open-source-
🤗 Model weights:
Skywork-OR1-Math-7B,Skywork-OR1-32B-Preview,Skywork-OR1-7B-Preview -
🤗 Training data:
Skywork-OR1-RL-Data -
🧑💻 Code:
Skywork-OR1 - We also release a Notion Blog to share detailed training recipes and extensive experimental results, analysis, and insights, dedicated to helping the community to better research, understand, and push the frontier of open reasoning models.
-
🤗 Model weights:
📖 Overview
The AIME24 scores versus training steps of Skywork-OR1-Math-7B in our multi-stage training pipeline.
The
Skywork-OR1
(Open Reasoner 1) model series consists of powerful math and code reasoning models trained using large-scale rule-based reinforcement learning with carefully designed datasets and training recipes. This series includes two general-purpose reasoning modelsl,
Skywork-OR1-7B-Preview
and
Skywork-OR1-32B-Preview
, along with a math-specialized model,
Skywork-OR1-Math-7B
.
-
Skywork-OR1-Math-7Bis specifically optimized for mathematical reasoning, scoring 69.8 on AIME24 and 52.3 on AIME25 — well ahead of all models of similar size. -
Skywork-OR1-32B-Previewdelivers the 671B-parameter Deepseek-R1 performance on math tasks (AIME24 and AIME25) and coding tasks (LiveCodeBench). -
Skywork-OR1-7B-Previewoutperforms all similarly sized models in both math and coding scenarios.
The final release version will be available in two weeks.
📊 Evaluation
We evaluate our models on AIME24, AIME25, and LiveCodeBench. Instead of using Pass@1, which is common in prior work, we introduce Avg@K as the primary metric. This metric robustly measures a model's average performance across K independent attempts, reducing the impact of randomness and enhancing the reliability of the results. We believe that Avg@K provides a better reflection of a model's stability and reasoning consistency.
We include the detailed results in the following table.
| Model | AIME24 (Avg@32) | AIME25 (Avg@32) | LiveCodeBench (8/1/24-2/1/25) (Avg@4) |
|---|---|---|---|
| DeepSeek-R1-Distill-Qwen-7B | 55.5 | 39.2 | 37.6 |
| Light-R1-7B-DS | 59.1 | 44.3 | 39.5 |
| DeepSeek-R1-Distill-Qwen-32B | 72.9 | 59.0 | 57.2 |
| TinyR1-32B-Preview | 78.1 | 65.3 | 61.6 |
| QwQ-32B | 79.5 | 65.3 | 61.6 |
| DeepSeek-R1 | 79.8 | 70.0 | 65.9 |
| Skywork-OR1-Math-7B | 69.8 | 52.3 | 43.6 |
| Skywork-OR1-7B-Preview | 63.6 | 45.8 | 43.9 |
| Skywork-OR1-32B-Preview | 79.7 | 69.0 | 63.9 |
⚙️ Training Recipe
We offer a brief overview of our data and training pipeline below. For more details, please refer to our Notion Blog here .
Data
- We select, clean, and curate a dataset of 110K verifiable, challenging, and diverse math problems and 14K coding questions from open-source datasets.
- We perform model-aware difficulty estimation for each problem and model and conduct rigorous quality assessment prior to training to ensure training efficiency and effectiveness.
Training
We develop a customized version of GRPO that leverages both data-wise and training-wise improvements:
- We perform both offline and online difficulty-based filtering and rejection sampling to improve training efficiency.
- We incorporate a multi-stage training pipeline coupled with adaptive entropy control and other techniques to enhance exploration and stability.
📄 Technical Report
Our technical report will be released soon. Stay tuned!
🙏 Acknowledgements
-
Both of our models are trained on top of
DeepSeek-R1-Distill-Qwen-7BandDeepSeek-R1-Distill-Qwen-32B. -
Both models are trained using
a custom fork
of the wonderful
verlproject.
📚 Citation
We will update the citation once the technical report is released. In the meantime, please cite the following:
@misc{skywork-or1-2025,
title={Skywork Open Reaonser Series},
author = {He, Jujie and Liu, Jiacai and Liu, Chris Yuhao and Yan, Rui and Wang, Chaojie and Cheng, Peng and Zhang, Xiaoyu and Zhang, Fuxiang and Xu, Jiacheng and Shen, Wei and Li, Siyuan and Zeng, Liang and Wei, Tianwen and Cheng, Cheng and Liu, Yang and Zhou, Yahui},
howpublished={\url{https://capricious-hydrogen-41c.notion.site/Skywork-Open-Reaonser-Series-1d0bc9ae823a80459b46c149e4f51680}},
note={Notion Blog},
year={2025}
}
Runs of Skywork Skywork-OR1-32B-Preview on huggingface.co
214
Total runs
0
24-hour runs
1
3-day runs
1
7-day runs
98
30-day runs
More Information About Skywork-OR1-32B-Preview huggingface.co Model
Url of Skywork-OR1-32B-Preview
Skywork-OR1-32B-Preview huggingface.co Url
Provider of Skywork-OR1-32B-Preview huggingface.co
Other API from Skywork
Updated:2024年10月25日
Total runs:
43.1K
Run Growth:
-33.4K
Growth Rate:
-77.56%
Total runs:
17.1K
Run Growth:
2.0K
Growth Rate:
11.49%
Total runs:
6.4K
Run Growth:
-6.0K
Growth Rate:
-94.11%
Total runs:
4.0K
Run Growth:
-8.0K
Growth Rate:
-200.13%
Total runs:
3.6K
Run Growth:
-1.2K
Growth Rate:
-33.26%
Total runs:
3.5K
Run Growth:
3.1K
Growth Rate:
86.41%
Total runs:
3.2K
Run Growth:
-10.3K
Growth Rate:
-324.12%
Total runs:
2.5K
Run Growth:
-4
Growth Rate:
-0.16%
Total runs:
1.5K
Run Growth:
106
Growth Rate:
7.24%
Total runs:
960
Run Growth:
155
Growth Rate:
16.15%
Total runs:
887
Run Growth:
-15.3K
Growth Rate:
-1721.53%
Total runs:
812
Run Growth:
176
Growth Rate:
21.67%
Total runs:
778
Run Growth:
33
Growth Rate:
4.24%
Total runs:
741
Run Growth:
399
Growth Rate:
53.85%
Total runs:
705
Run Growth:
332
Growth Rate:
47.09%
Total runs:
611
Run Growth:
-63
Growth Rate:
-10.31%
Total runs:
598
Run Growth:
123
Growth Rate:
20.57%
Total runs:
500
Run Growth:
-408
Growth Rate:
-81.60%
Total runs:
471
Run Growth:
129
Growth Rate:
27.39%
Total runs:
346
Run Growth:
-1.5K
Growth Rate:
-447.40%
Total runs:
335
Run Growth:
96
Growth Rate:
28.66%
Total runs:
238
Run Growth:
-552
Growth Rate:
-231.93%
Total runs:
220
Run Growth:
13
Growth Rate:
5.91%
Total runs:
210
Run Growth:
-27
Growth Rate:
-12.86%
Total runs:
163
Run Growth:
15
Growth Rate:
9.20%
Total runs:
127
Run Growth:
-26
Growth Rate:
-20.47%
Total runs:
61
Run Growth:
-9
Growth Rate:
-14.75%
Total runs:
59
Run Growth:
59
Growth Rate:
100.00%
Total runs:
56
Run Growth:
-48
Growth Rate:
-85.71%
Total runs:
54
Run Growth:
-19
Growth Rate:
-35.19%
Total runs:
44
Run Growth:
-25
Growth Rate:
-56.82%
Total runs:
42
Run Growth:
-50
Growth Rate:
-119.05%
Total runs:
37
Run Growth:
9
Growth Rate:
24.32%
Total runs:
33
Run Growth:
6
Growth Rate:
18.18%
Total runs:
33
Run Growth:
6
Growth Rate:
18.18%
Total runs:
32
Run Growth:
9
Growth Rate:
28.13%
Total runs:
24
Run Growth:
-7
Growth Rate:
-29.17%
Total runs:
22
Run Growth:
-9
Growth Rate:
-40.91%
Total runs:
21
Run Growth:
0
Growth Rate:
0.00%
Total runs:
18
Run Growth:
1
Growth Rate:
5.56%
Total runs:
18
Run Growth:
-6
Growth Rate:
-33.33%
Total runs:
18
Run Growth:
-20
Growth Rate:
-111.11%
Total runs:
16
Run Growth:
6
Growth Rate:
37.50%
Total runs:
14
Run Growth:
-17
Growth Rate:
-121.43%
Total runs:
0
Run Growth:
0
Growth Rate:
0.00%