Pakistan Unveils First Urdu AI Models: Qalb, ALIF & National LLM

Pakistan has taken a major step in artificial intelligence with the launch of several large language models designed specifically for Urdu. These initiatives aim to make AI accessible to the country’s 230 million Urdu-speaking citizens and reduce reliance on foreign technology. Both private teams and government bodies are building systems that understand, generate, and interact in Urdu, creating new opportunities in education, public services, and local content development.

Qalb: Pakistan’s Homegrown Urdu Model

In January 2026, a student-led team introduced Qalb, meaning “heart” in Urdu. Developed by Hassan alongside Jawad Ahmed and Muhammad Awais, it is regarded as Pakistan’s first fully homegrown Urdu AI model.

The multi-billion-parameter system was trained on billions of Urdu words and phrases, enabling it to generate culturally relevant language. Its creators say Qalb will support students, researchers, and content creators, helping bridge the gap between advanced AI tools and Urdu-speaking communities.

ALIF: A Local Version of Global AI

Earlier, in October 2025, Pakistan’s Ministry of IT and Telecommunication, in partnership with Meta, launched ALIF. Built on Meta’s Llama technology, ALIF has been fine-tuned with local datasets to better understand Urdu. The assistant is intended for public-sector use, including government offices and community initiatives. Officials describe it as a practical blend of global AI expertise and local linguistic insight to enhance accessibility and efficiency.

National LLM for Government Services

A separate state-backed project involving the National University of Sciences and Technology and the National Information Technology Board is developing a national LLM. It is being trained on more than two million locally sourced documents and will be integrated into services such as healthcare, education, and administration.

Officials say the initiative forms part of a broader plan to digitize public services while reducing dependence on foreign AI platforms. Once launched, the model could significantly expand access to information and essential services for Urdu-speaking citizens across Pakistan. This transformative push highlights Pakistan’s commitment to fostering indigenous innovation and ensuring that language is no barrier to technological participation in an increasingly digital world for its people, nationwide digital progress.