Text model

Ulam-1-Small

Ulam-1-Small is a 3.086B-parameter open-weight language model from Ulam AI, released under the MIT license and built on a Qwen2.5-Coder/VibeThinker lineage. It targets exploratory mathematical problem solving and proof-style reasoning, using explicit step-by-step reasoning delimiters and a 131,072-token context window. It is not a theorem-certification system, so its proofs and claims require independent expert review.