QCRI

The Qatar Computing Research Institute (QCRI) is a national research institute in Doha, Qatar, established in 2010 by the Qatar Foundation and now operating as one of the four research institutes of Hamad Bin Khalifa University (HBKU). Founded as part of Qatar's broader National Research Strategy, QCRI conducts research across Arabic-language natural-language processing, computational social science, distributed systems, computer vision, cybersecurity, and AI for social good. The institute's flagship public-facing AI release is Fanar, an Arabic-and-English large language model launched December 2024 that the institute developed jointly with the Qatar Ministry of Communications and Information Technology and that has been positioned as Qatar's contribution to a sovereign Arabic-language AI ecosystem. As of April 2026, QCRI is one of the principal Gulf-region academic AI research institutes alongside MBZUAI (Abu Dhabi) and KAUST (Saudi Arabia), with substantial Arabic-NLP research output and government-aligned strategic positioning.

At a glance

• Founded: 2010 in Doha, Qatar, by the Qatar Foundation. Operates as one of four research institutes of Hamad Bin Khalifa University since HBKU's establishment.
• Status: Public-funded national research institute under Hamad Bin Khalifa University and the Qatar Foundation.
• Funding: Public-sector funding from the Qatar Foundation and the Qatar government. Specific institutional budget figures are not publicly disclosed.
• Executive Director: Ahmed Elmagarmid, Executive Director of QCRI since the institute's founding period. Long-tenured computer-science professor; former Purdue University professor and director of the Cyber Center at Purdue.
• Other notable leadership: Dr. Mounir Hamdi, Founding Dean of the College of Science and Engineering at HBKU and longtime collaborator with QCRI. Senior research staff distributed across the institute's research groups including Dr. Ahmed Ali (Arabic speech and NLP), Dr. Wajdi Zaghouani (Arabic computational linguistics), and Dr. Marco Brambilla (data science).
• Open weights: Yes, partial. The Fanar Arabic-English large language model was released with both open-weights research variants and closed commercial-API variants. Selected adjacent research releases have been open-source.
• Flagship outputs: Fanar Arabic-English large language model (December 2024), the Aamal Arabic ASR system, the Qatar Computing Research Institute Arabic NLP Toolkit, the AraT5 Arabic text-to-text model, and a long-running publication record at ACL, EMNLP, NAACL, and adjacent NLP and computer-science venues.

Origins

QCRI was established in 2010 by the Qatar Foundation as a national research institute focused on computing and information sciences. The founding year aligned with Qatar's broader 2008-launched National Research Strategy, which committed to a target of 2.8 percent of GDP for research and development and identified information technology as one of four priority research pillars (alongside energy and environment, health, and social sciences). The institute was designed from inception as a peer to comparable national-strategy AI research institutes elsewhere, with the explicit positioning of building Qatar's domestic AI research capability and Qatar-language and Qatar-context research outputs (Arabic NLP, computational social science of the MENA region, smart-city research grounded in Doha's developing infrastructure).

Ahmed Elmagarmid was recruited from Purdue University to serve as the founding Executive Director and has continued in the role across the institute's history. Through the 2010s, QCRI built research depth across what became four principal research groups: Arabic NLP and speech, computational social science, distributed systems and data engineering, and AI for social good (including disaster response and humanitarian-applications research). The institute's affiliation with Hamad Bin Khalifa University, established in 2010 as the Qatar Foundation's research-and-graduate university, anchored an academic-credentialing structure that allowed QCRI to host PhD students, postdoctoral researchers, and faculty appointments alongside its research-staff scientists.

The Arabic NLP research line has been the institute's most internationally visible output. The institute's Arabic NLP Toolkit (Farasa for morphological analysis, segmentation, and POS tagging) has been used widely in academic Arabic-NLP research. The AraT5 Arabic text-to-text model, released open-weights in 2022, was among the principal Arabic-language research releases of its period. The 2010s also produced substantial Arabic speech-recognition research output through the Aamal program.

The December 2024 launch of Fanar marked the institute's principal public-facing generative-AI release. Fanar was developed jointly between QCRI and the Qatar Ministry of Communications and Information Technology with the explicit positioning as a sovereign Arabic-and-English large language model. Industry coverage has reported Fanar's training corpus combining Arabic web data, Qatar-government-aligned content, and English-language general-purpose data, with the model being released through a Qatar-domiciled Fanar API platform alongside selected open-weights research variants.

The 2024 to 2026 period has continued the Fanar iteration alongside the institute's broader research output, with substantial cooperation reported with international academic peers on Arabic-NLP benchmark development and computational-social-science research.

Mission and strategy

QCRI's stated mission is to advance computing and information-sciences research in service of Qatar's national priorities, with research output spanning foundational AI research, applied research with Qatar-government and Qatar-industry partners, and capacity building for Qatar's domestic computing-research talent pipeline. The strategy combines three threads. First, Arabic-language NLP and speech research that positions Qatar as a contributor to the Arabic-AI ecosystem alongside peer institutions in the broader MENA region. Second, applied-AI research with Qatar-government and Qatar-industry partners, with the December 2024 Fanar launch as the principal public-facing example. Third, computational-social-science and AI-for-social-good research that reflects Qatar's positioning as a regional research-and-policy hub.

The competitive premise is that Arabic-language and MENA-context research requires sustained sovereign-state investment that international academic and commercial AI organizations have historically under-resourced, and that Qatar can produce structurally relevant AI capabilities (Arabic NLP, MENA computational social science) that will not emerge without that investment.

Models and products

• Fanar. Arabic-English large language model. Launched December 2024 jointly with the Qatar Ministry of Communications and Information Technology. Released through both open-weights research variants and a closed-API commercial offering through the Fanar platform.
• AraT5 and earlier Arabic-language models. The principal pre-Fanar Arabic-NLP research releases. Open-weights distribution.
• Farasa Arabic NLP Toolkit. Long-running open-source Arabic-text-processing toolkit. Used in academic Arabic-NLP research globally.
• Aamal Arabic ASR. Arabic speech-recognition research and tooling.
• Active publication record. At ACL, EMNLP, NAACL, COLING, ICASSP, and adjacent NLP, speech-recognition, and computer-science venues. Research areas span Arabic NLP, computational social science, distributed systems, and computer vision.
• AI for Social Good research. Including computational-disaster-response, humanitarian-applications, and adjacent research lines.

Distribution channels include the Fanar API platform, Hugging Face for open-weights model releases, and academic publication for the broader research output.

Benchmarks and standing

QCRI's evaluation framework focuses on Arabic-NLP benchmark performance, academic publication impact, and the practical deployment of QCRI-developed AI capabilities into Qatar government and Qatar industry. Fanar has been benchmarked against international Arabic-language alternatives including Cohere's Aya, Microsoft's Jais (developed with G42 and MBZUAI), and OpenAI's GPT-4o on Arabic-language tasks, with QCRI-published benchmark results indicating competitive Arabic-language performance.

Industry coverage has consistently characterized QCRI as one of the principal Gulf-region academic AI research institutes, with the December 2024 Fanar launch and the long-running Arabic-NLP research output as principal validating data points.

Leadership

As of April 2026, QCRI's senior leadership includes:

• Ahmed Elmagarmid, Executive Director.
• Senior faculty leadership across the Arabic NLP and speech, computational social science, distributed systems, and AI for social good research groups.

The senior leadership has been comparatively stable across the institute's operating history, with Elmagarmid continuing as Executive Director from the founding period.

Funding and backers

Public-sector funding from the Qatar Foundation and the Qatar government. Specific institutional budget figures are not publicly disclosed. The institute's affiliation with Hamad Bin Khalifa University provides academic-program funding alongside QCRI's research-program budget.

Industry position

QCRI occupies a distinctive position as the principal Qatar academic AI research institute, with the Fanar Arabic-English large language model, the long-running Arabic-NLP research output, and the institute's role within Qatar's broader National Research Strategy. Industry coverage has consistently grouped QCRI with MBZUAI, KAUST, TII, and the Saudi-led HUMAIN initiative as the principal Gulf-region AI research and development organizations. Among that group, QCRI's distinguishing positioning is the long operating history (since 2010, the earliest of the cohort) and the explicit academic-research orientation that contrasts with the more commercial positioning of TII and HUMAIN.

Competitive landscape

• MBZUAI (Abu Dhabi). Direct Gulf-region academic AI research institute. Different governance (graduate-research university focus); smaller research-staff size.
• KAUST AI Initiative (Saudi Arabia). Saudi academic AI research peer.
• TII (Abu Dhabi). UAE state-affiliated commercial-AI research organization (developer of Falcon).
• G42, Core42, Inception, HUMAIN. Gulf-region commercial AI organizations operating in adjacent commercial-and-state spaces.
• Cohere for AI, AI4Bharat, Sarvam AI, Masakhane. Multilingual-AI peer organizations focused on under-served languages outside the Western frontier.
• Stanford NLP, MIT CSAIL, CMU SCS. Western academic NLP and AI research peers; QCRI maintains active research collaborations with several of these.

Outlook

• Continued Fanar iteration and any successor model release.
• Continued Arabic-NLP research output and benchmark contributions.
• The continued Qatar government and industry partnership engagement.
• The competitive dynamic with MBZUAI and the broader Gulf-region academic AI ecosystem.
• Continued academic publication output at ACL, EMNLP, NAACL, and adjacent venues.

Sources

• QCRI official site. Institute reference.
• Hamad Bin Khalifa University. Parent university.
• Qatar Foundation. Founding and continuing funder.
• Fanar. Arabic-English large language model.
• Ahmed Elmagarmid LinkedIn. Executive Director reference.