Complex math reasoning dataset

What’s in this dataset:

• University-level math problems with step-by-step solutions*

• Format: LaTeX formulas with comments

• Multiple solutions provided for ~10% 

of the problems 

  *natural language step-by-step explanations can be added by request

Dataset design

Real-world non-synthetic mathematical problems are produced by our partner, a US-based EdTech company. 



Solutions are validated and tested by the internal team and external subject matter experts from universities to ensure educational relevance and rigor.