The weights and data pipeline are open sourced and described explicitly in the paper they published. The non-reasoning data isn't nearly as interesting as the r

The weights and data pipeline are open sourced and described explicitly in the paper they published. The non-reasoning data isn't nearly as interesting as the reasoning data though