OctoThinker/OctoThinker-3B-Short-Base

OctoThinker: Mid-training Incentivizes Reinforcement Learning Scaling

OctoThinker-3B-Short-Base

The OctoThinker family is built on carefully studied mid-training insights, starting from the Llama-3 family, to create a reinforcement learning–friendly base language model.

Training Recipe

<div style="display: flex; justify-content: left; gap: 20px;"> <img src="https://cdn-uploads.huggingface.co/production/uploads/62cbeb2d72dfd24b86bdf977/2sFzePngjjopTs0SeCS9R.png" alt="Data Pipeline" style="width:90%;"> </div>

Evaluation Results

Note that we adopt the few-shot prompting evaluation for these base language models.

<div style="display: flex; justify-content: left; gap: 20px;"> <img src="https://cdn-uploads.huggingface.co/production/uploads/62cbeb2d72dfd24b86bdf977/UCZ9MahRYqLY0iKjiWMqS.png" alt="Data Pipeline" style="width:80%;">

</div>

More about OctoThinker

<div style="display: flex; justify-content: left; gap: 20px;"> <img src="https://cdn-uploads.huggingface.co/production/uploads/62cbeb2d72dfd24b86bdf977/bn85CEB_DW6azJ7KJp11Q.png" alt="Data Pipeline" style="width:100%;"> </div>

Citation

Check out our paper for more details. If you use our models, datasets or find our work useful, please cite

@article{wang2025octothinker,
  title={OctoThinker: Mid-training Incentivizes Reinforcement Learning Scaling},
  author={Wang, Zengzhi and Zhou, Fan and Li, Xuefeng and Liu, Pengfei},
  year={2025},
  journal={arXiv preprint arXiv:2506.20512},
  note={Preprint}
}