GigaChat Family: Efficient Russian Language Modeling Through Mixture of Experts Architecture
This addresses the need for efficient Russian language models for NLP research and industrial applications, though it is incremental as it applies an existing mixture of experts architecture to a new language.
The paper tackles the limited development of foundational large language models for Russian by introducing the GigaChat family, which achieves competitive performance on Russian and English benchmarks while being made available via API, Telegram bot, and open-source release.
Generative large language models (LLMs) have become crucial for modern NLP research and applications across various languages. However, the development of foundational models specifically tailored to the Russian language has been limited, primarily due to the significant computational resources required. This paper introduces the GigaChat family of Russian LLMs, available in various sizes, including base models and instruction-tuned versions. We provide a detailed report on the model architecture, pre-training process, and experiments to guide design choices. In addition, we evaluate their performance on Russian and English benchmarks and compare GigaChat with multilingual analogs. The paper presents a system demonstration of the top-performing models accessible via an API, a Telegram bot, and a Web interface. Furthermore, we have released three open GigaChat models in open-source (https://huggingface.co/ai-sage), aiming to expand NLP research opportunities and support the development of industrial solutions for the Russian language.