New Large Language Model Released by Yandex

Yandex has released a new large language model (LLM) titled “AliceAI-Foundation-80B-A3B-Base,” according to technical documentation hosted on Hugging Face. The model is positioned by developers as a Russian-developed competitor to existing industry benchmarks, specifically targeting the performance levels of Qwen 35B and DeepSeek V4 Flash.

Technical Specifications and Architecture

According to reports circulating on Reddit, the AliceAI-Foundation-80B-A3B-Base utilizes a fully custom architecture. Technical discussions suggest that the model is not a fine-tuned version of the Qwen series, nor does it rely on the Llama architecture.

Notably, the model is currently a base model and has not undergone post-training, distinguishing it from recent iterations of Qwen models such as Qwen 3.5 or 3.6. Users have noted that, at the time of release, the model lacks support for Llama.cpp. While the model’s performance is being discussed in developer communities, independent verification of its benchmarking results against its stated competitors has not yet been established.

Market Context and Reception

The release comes amid a complex geopolitical landscape for the technology sector. Discussions on social media platforms have highlighted the controversial nature of the release, with some users pointing to the company’s international presence and the current geopolitical tensions involving Russia.

While some observers in developer forums have suggested the model may offer competitive advantages over certain European AI developments, the reception remains polarized. Discussions on Reddit reflect a divide between those interested in the technical implications of a new custom architecture and those focused on the political implications of Russian-developed AI technology.

As the global AI race continues, the entry of a custom-architected 80B model into the open-source ecosystem marks a specific development in the efforts of non-Western firms to provide alternatives to models developed by major US-based entities such as Alphabet Inc. or Meta Platforms Inc. However, the practical utility and widespread adoption of the AliceAI-Foundation-80B-A3B-Base remain to be seen as the developer community evaluates its performance and integration capabilities.