Abu Dhabi, June 25 (WAM) - Khalifa University for Science and Technology announced the launch of "TelecomGPT-R1", the latest AI model developed by the university's Institute for Digital Future to serve the telecommunications sector.

The model, launched on June 22, topped the "Open Telco Leaderboard" issued by the Global System for Mobile Communications Association (GSMA), the leading standard in the industry for evaluating the performance of AI models in real-world telecommunications tasks.

The model was showcased at the Mobile World Congress Shanghai 2026 and outperformed all other evaluated models, including prominent closed-source systems developed by major global technology companies.

"TelecomGPT-R1" achieved a performance rate of 89.6% on the leaderboard, the highest among all tested models, whether open-source or closed-source, general-purpose, or tailored for the telecommunications sector.

The GSMA manages this leaderboard, a global organization that unites the mobile communications ecosystem and includes most mobile network operators worldwide. This leaderboard is the widely adopted standard for measuring how well AI models can meet the technical requirements of actual communication networks.

His Excellency Professor Ibrahim Al Hagry, President of Khalifa University, said: "TelecomGPT-R1 represents a model of what Khalifa University aims to achieve by linking advanced scientific research with the future needs of vital sectors. Such achievements not only contribute to developing next-generation technologies such as 6G, but also enhance the university’s role in producing knowledge, innovation, and building national capabilities in strategic priority areas for the UAE."

He added: "High-quality research achievements are a key factor in establishing universities' global standing and enhancing their academic competitiveness. This achievement comes at a time when Khalifa University has achieved its highest ever ranking in the QS World University Rankings 2027, moving up 30 places to 147th globally, confirming the strength of its research system, the wide impact of its scientific work, and its effective partnerships with the industrial sector."

Luis Paul, Director of AI Technologies at the GSM Association, said that developing customized models for the telecommunications sector characterized by accuracy, efficiency, and reliability is a fundamental step in expanding the application of artificial intelligence in the sector. He congratulated Khalifa University on this achievement.

The model was developed from scratch to be tailored specifically for the telecommunications sector, combining technical linguistic understanding, network-related reasoning, and operational knowledge within an integrated system.

Unlike general-purpose models, which often struggle with the high specialized knowledge of this sector, TelecomGPT-R1 is designed to support tasks such as network troubleshooting and repair, fault analysis, and assisting in deployment processes – capabilities that the telecommunications industry will increasingly rely on as networks move towards greater automation and AI reliance.

The open-source model allows telecom network operators, technology providers, and researchers to examine, test, and adapt it for their own systems.

"TelecomGPT-R1 demonstrates that open AI models in the telecommunications sector can lead this field rather than just participate in it," said Professor Marwan Dabbah, Director of the Future Digital Institute at Khalifa University. "Leading the Open Telco Leaderboard by GSMA is an important milestone for Khalifa University and the Future Digital Institute, reflecting our commitment to building transparent, relevant AI systems that meet the needs of the telecommunications sector for the future of communications."

TelecomGPT-R1 complements a broader research program led by Khalifa University and the Future Digital Institute focusing on next-generation network AI, including customized language models for the telecom sector, foundational models for radio frequencies, and AI systems tailored for technical standard-related inference.