Clear answers about Khazri

Khazri is a large language model family built for Azerbaijani and local context. Its purpose is to strengthen Azerbaijan's AI ecosystem and reduce dependence on foreign infrastructure.

The Khazri team presents the project as Azerbaijan's first local large language model family. Khazri 1 is its first generation; Khazri 2 and Khazri 2 Mini are the newer models published on Hugging Face.

The current flagship is the 250M-parameter Khazri 2. It uses modern attention and activation methods together with the SYNAPSE architecture.

The 100M-parameter Khazri 2 Mini is published on Hugging Face for applications that need less compute, lower latency and more compact deployment.

Yes. The GPT-2-based Khazri 1 is available on Hugging Face under GPL-3.0. Its capabilities are substantially more limited than those of the new Khazri 2 family.

Khazri 2 and Khazri 2 Mini are published as open weights on Hugging Face. Contact our team via WhatsApp or contact@khazri.dev to discuss integration, deployment and commercial collaboration.

SYNAPSE is the name of the architecture developed by the Khazri team for the Khazri 2 family. We currently share its name and general direction; not every technical detail has been published.

Priority areas include Azerbaijani citizen services, enterprise knowledge assistants, document processing, search, summarisation, education and cultural content. Exact suitability is assessed before integration.

The benchmark visual shows internal results from a selected preview test configuration supplied by the Khazri team. They are not an independent certification and may change before final release.

The Khazri model belongs to Softyu.