DR-Tulu is the first open deep research model directly trained for open-ended, long-form research using Reinforcement Learning with Evolving Rubrics (RLER).