1. Tutorial Introduction

The DeepCoder-14B-Preview project is a 14B coding model released by AGENTICA on April 8, 2025. The model is fine-tuned from DeepSeek-R1-Distilled-Qwen-14B LLM for code reasoning, using distributed reinforcement learning (RL) to scale to long context lengths. The model achieved a Pass@1 accuracy of 60.6% on LiveCodeBench v5 (8/1/24-2/1/25), an improvement of 8% over the base model (53%), and achieved similar performance to OpenAI's o3-mini with only 14B parameters.

This tutorial uses the DeepCoder-14B-Preview model as a demonstration case and uses the 8-bit quantization method provided by bitsandbytes to optimize video memory usage. It should be noted that this quantization method may have a certain impact on the performance of the model. The computing power resource uses RTX4090.

2. Operation steps

1. After starting the container, click the API address to enter the Web interface

If "Bad Gateway" is displayed, it means the model is initializing. Please wait for about 1-2 minutes and refresh the page.

2. Functional Demonstration

If "Error: HTTPConnectionPool(host="localhost, port=8000): Max retries exceeded with url: /v1/completions" is displayed as shown below, it means the model is initializing. Please wait for about 1-2 minutes and refresh the page.

Citation Information

Thanks to GitHub user SuperYang  For the production of this tutorial, the project reference information is as follows:

@misc{deepcoder2025,
  title={DeepCoder: A Fully Open-Source 14B Coder at O3-mini Level},
  author={Michael Luo, Sijun Tan, Roy Huang, Xiaoxiang Shi, Rachel Xin, Colin Cai, Ameen Patel, Alpay Ariyak, Qingyang Wu, Ce Zhang, Li Erran Li, Raluca Ada Popa, Ion Stoica},
  howpublished={\url{https://pretty-radio-b75.notion.site/DeepCoder-A-Fully-Open-Source-14B-Coder-at-O3-mini-Level-1cf81902c14680b3bee5eb349a512a51}},
  note={Notion Blog},
  year={2025}
}

@misc{deepscaler2025,
  title={DeepScaleR: Surpassing O1-Preview with a 1.5B Model by Scaling RL},
  author={Michael Luo and Sijun Tan and Justin Wong and Xiaoxiang Shi and William Y. Tang and Manan Roongta and Colin Cai and Jeffrey Luo and Li Erran Li and Raluca Ada Popa and Ion Stoica},
  year={2025},
  howpublished={\url{https://pretty-radio-b75.notion.site/DeepScaleR-Surpassing-O1-Preview-with-a-1-5B-Model-by-Scaling-RL-19681902c1468005bed8ca303013a4e2}},
  note={Notion Blog}
  year={2025}
}

Exchange and discussion

🖌️ If you see a high-quality project, please leave a message in the background to recommend it! In addition, we have also established a tutorial exchange group. Welcome friends to scan the QR code and remark [SD Tutorial] to join the group to discuss various technical issues and share application effects↓