T-pro 2.0: An Efficient Russian Hybrid-Reasoning Model and Playground Paper β’ 2512.10430 β’ Published Dec 11, 2025 β’ 113