Models we've trained and released, with links to papers, weights, and inference notes.
A 24B-parameter reasoning model post-trained for chemistry.
