Yandex has developed a large language model, Alice AI Foundation LLM, trained from scratch without third-party foreign components, with 80 billion parameters and about 3 billion active simultaneously, using 18 trillion tokens. The company opened the model's weights, allowing developers to run it independently and fine-tune it for their tasks, under a license permitting commercial and research use. Yandex expects to use it as a basis for a future system with reasoning capabilities and agent functions for Alice AI. However, it is too early to call it an officially recognized sovereign model; Yandex positions it as fully sovereign, but the law establishes a separate compliance procedure, with main norms coming into force on March 1, 2027. A sovereign large fundamental model must be developed by a Russian legal entity with a technically reproducible life cycle, Russian data centers for user responses and data storage, and pass compliance confirmation; national model requirements are softer. In July, Sber's CEO German Gref said only the bank had a fully domestic large language model and Yandex was allegedly fine-tuning Chinese Qwen neural networks from Alibaba, which Yandex denied. A Sber representative called the emergence of Alice AI Foundation evidence of the high level of Russian engineering teams, but noted many models with up to 100 billion parameters exist, and the country's global position is determined by flagship models with hundreds of billions and trillions of parameters. Experts consider it part of Yandex-Sber competition