Pakistan has launched its first homegrown Urdu-language AI model, Qalb, developed by Taimoor Hassan, a Pakistani student currently studying in the United States. Qalb is being hailed as the world’s largest large language model built exclusively for Urdu, trained on 1.97 billion tokens and evaluated across seven international benchmarking frameworks.
According to its developers, Qalb surpasses existing Urdu-focused AI models in key real-world performance metrics, marking a significant step in advancing AI for the national language.
Taimoor Hassan, a young entrepreneur and former Microsoft Cup winner, has launched and exited 13 startups and represented Pakistan on multiple international platforms.
He completed his Bachelor’s degree in Computer Science from FAST University and is now pursuing a Master’s in Computer Science and Software Engineering at Auburn University, where he continues to work on advanced AI and technology projects.
With over 230 million Urdu speakers globally, Qalb is designed to bridge the technology gap by providing AI capabilities in Urdu for education, digital services, startups, local businesses, and voice-based AI agents.
The model is expected to support the development of localized solutions and applications that cater specifically to Urdu-speaking audiences.
Hassan explained that the project is part of a broader mission to contribute to Pakistan’s technological future. He collaborated closely with his undergraduate teammates, Jawad Ahmed and Muhammad Awais, and highlighted plans to refine and expand localized AI models for niche industries.
The team emphasized that innovation is no longer limited to large corporations. Hassan noted that modern technology allows small, agile teams to build tools that educate, automate, and serve millions of people without the need for massive budgets.
Qalb exemplifies this approach by offering advanced AI capabilities in a language that has long been underserved by cutting-edge technology. The model positions Pakistan at the forefront of localized AI development, opening doors for more accessible, inclusive, and culturally relevant AI solutions.
