The article introduces PolitNuggets, a new benchmark for evaluating the ability of large reasoning models to synthesize political facts from diverse sources. It highlights the challenges these models face in retrieving fine-grained details and emphasizes the need for improvements in efficiency and multilingual capabilities.