This is a preliminary implementation of the paper "Xolver: Multi-Agent Reasoning with Holistic Experience Learning Just Like an Olympiad Team". More tasks and settings will be released soon. You may see some additional Xolver logs here.

Md Tanzib Hosain, Salman Rahman, Md Kishor Morol, Md Rizwan Parvez

To clone the project, run:

git clone https://github.com/kraritt/Xolver && cd Xolver

The code for running GSM, AIME, MATH and LiveCodeBench tasks may be found in the following subfolders

• ./gsm/ contains code for running GSM
• ./aime/ contains code for running AIME
• ./math/ contains code for running MATH
• ./lcb/ contains code for running LiveCodeBench results

To prepare a new conda or python virtual environment, run:

pip install -r requirements.txt

GSM:

To generate and evaluated answer for GSM problems through Xolver, run:

cd ./gsm python gsm.py

You can download the GSM dataset here

AIME:

To generate and evaluated answer for AIME problems through Xolver, run:

cd ./aime python aime.py

You can download the AIME datasets here and here

MATH:

To generate and evaluated answer for MATH problems through Xolver, run:

cd ./math python math.py

You can download the MATH dataset here

LiveCodeBench:

To generate and evaluated answer for LiveCodeBench problems through Xolver, run:

cd ./lcb python lcb.py

You can download the LiveCodeBench dataset here

You can download the math external retrieval corpus here and the code external retrieval corpus here

If you would like to cite the paper, here is a bibtex file:

@article{hosain2025xolver,
      title={𝕏olver: Multi-Agent Reasoning with Holistic Experience Learning Just Like an Olympiad Team}, 
      author={Md Tanzib Hosain and Salman Rahman and Md Kishor Morol and Md Rizwan Parvez},
      journal={arXiv preprint},
      year={2025}
}