T-pro 2.0: An Efficient Russian Hybrid-Reasoning Model and Playground

TL;DR

T-pro 2.0 is an open Russian LLM designed for hybrid reasoning and efficient inference, featuring a Cyrillic tokenizer, a speculative-decoding pipeline, and comprehensive resources for research and application development.

Contribution

It introduces T-pro 2.0, a novel open-weight Russian language model with optimized inference and reasoning capabilities, along with extensive resources for reproducible research.

Findings

Reduced inference latency through EAGLE speculative decoding

Enabled reasoning trace generation and direct answering

Released comprehensive datasets and model weights for research

Abstract

We introduce T-pro 2.0, an open-weight Russian LLM for hybrid reasoning and efficient inference. The model supports direct answering and reasoning-trace generation, using a Cyrillic-dense tokenizer and an adapted EAGLE speculative-decoding pipeline to reduce latency. To enable reproducible and extensible research, we release the model weights, the T-Wix 500k instruction corpus, the T-Math reasoning benchmark, and the EAGLE weights on Hugging Face. These resources allow users to study Russian-language reasoning and to extend or adapt both the model and the inference pipeline. A public web demo exposes reasoning and non-reasoning modes and illustrates the speedups achieved by our inference stack across domains. T-pro 2.0 thus serves as an accessible open system for building and evaluating efficient, practical Russian LLM applications.