TII Abu Dhabi Launches Arabian Falcon-H1, Building the World’s Leading Arabian AI Model

Monday 05 January 2026 17:03
| Updated:

Monday 05 January 2026 17:04

Share

• 
  Share on Facebook
• 
  Share on Twitter
• 
  Share on LinkedIn
• 
  Share on WhatsApp
• 
  Share on Email

The Technological Innovation Institute (TII), the applied research arm of Abu Dhabi’s Advanced Technology Research Council (ATRC), has announced the Arabian Falcon-H1, a newly developed large language model built on a hybrid Mamba-Transformer architecture. Representing a complete change from the previous transformer-based version, this new model establishes itself as the highest performing system on the Open LLM Arabic Language (OALL) Leaderboard.

This press release features multimedia. View the full release here:

This achievement makes Falcon-H1 Arabic the leading Arabic AI model available today, outperforming models several times larger, while providing advanced accuracy, context handling, and linguistic representation.

His Excellency Faisal al Bannai, Advisor to the President of the UAE and Secretary General of the Advanced Technology Research Council said“Falcon-H1 Arab reflects our continued commitment to strengthening the UAE’s position as a global hub for advanced technology and responsible AI. By delivering models that support the linguistic and cultural needs of the region, we are enabling innovation that is accessible, relevant and impactful across our society. This achievement is a testament to the depth of research talent and expertise at TII.”

Based on the strong reception of the Falcon-Arabic model released earlier this year, which highlighted the community’s clear need for a high-quality Arabic LLM, TII has advanced its work with the new Falcon-H1 Arabic family. Available in 3B, 7B, and 34B parameter sizes, this model is designed to meet a variety of infrastructure needs and use cases. Falcon-H1 Arabic introduces improvements in data quality, dialect coverage, long-context stability, and mathematical reasoning, enabling more accurate, reliable, and context-aware understanding of Arabic across real-world applications.

Said Dr Najwa Aaraj, CEO of TII“The development of Falcon-H1 Arab builds on years of foundational work in Arabic AI and responds directly to the needs of our community, including developers and businesses. By advancing architecture, data quality, and long-context reasoning, we are creating technology that opens up new possibilities in education, healthcare, governance, and enterprise, and more, all in Arabic. This model represents an important step in our mission to deliver world-class AI that serves the region and contributes to global progress.”

Benchmark Results

On the OALL leaderboard – which evaluates models in a variety of Arabic comprehension and reasoning tasks – Falcon-H1 Arabic shows clear performance leadership:

• The 3B model scored an average of 61.87%, 10 points ahead of leading 4B competitors, such as Microsoft’s Phi-4 Mini.

• The 7B model scored an average of 71.47%, outperforming all ~10B models, including the Fanar-1-9B model from Qatar and the HUMAIN ALLaM 7B model from Saudi Arabia.

• The 34B model scored 75.36%, even outperforming 70B+ parameter systems, including China’s Qwen2.5 72B and META’s Llama-3.3 70B.

In addition to OALL, the Falcon-H1 Arabic language model also achieved outstanding results on more targeted benchmarks, including, (i) 3LM, for STEM reasoning, (ii) ArabCulture, for cultural and contextual understanding, and (iii) AraDice (dialect understanding).

Overall, these results represent a breakthrough moment for Arab AI. Falcon-H1 Arabic not only outperforms models several times larger, in both general and specific benchmarks, but also exhibits levels of linguistic depth, reasoning ability, and efficiency that are new benchmarks in the field. This makes the Falcon-H1 Arabic the most capable and versatile Arabic model developed to date.

Dr Hakim Hacid, Head Researcher of the TII Artificial Intelligence and Digital Research Center (AIDRC) said“This model reflects our focus on building Arabic AI that is not only more advanced, but also truly useful in the real world. By increasing efficiency, depth of understanding, and language coverage, we are enabling AI systems that can better support institutions, developers, and communities across the region.”

The model also expands the context significantly, with a window of up to 256K tokens, allowing the model to work on a large amount of information in a single interaction. In practice, this means users can, for example, analyze long legal documents, medical records, academic papers, or company knowledge bases without losing context or continuity – a capability that was previously impossible at this scale.

TII’s Falcon AI model has ranked number one across regional and global benchmarks since 2023, with the Arabian Falcon-H1 now leading the LLM Arabian Open Leaderboard across model sizes. These results demonstrate TII’s ability to build a sovereign AI capability that competes at the highest global level, while advancing Abu Dhabi and the UAE’s leadership in Arab AI research and innovation.

The new models are now accessible within the playground via TII’s public interface at:

Source: AETOS Wire

View source version on businesswire.com:

Contact

Warrayat Women

jinan.warrayat@tii.ae

Abstract

TII Abu Dhabi Launches Arabian Falcon-H1, Building the World’s Leading Arabian AI Model

VIDEO: TII Abu Dhabi Launches Arabian Falcon-H1, Building World’s Leading Arabian AI Model (Video: AETOSWire)

TII Abu Dhabi Launches Arabian Falcon-H1, Building World’s Leading Arabian AI Model (Graphic: AETOSWire)

TII Abu Dhabi Launches Arabian Falcon-H1, Building World’s Leading Arabian AI Model (Graphic: AETOSWire)

TII Abu Dhabi Launches Arabian Falcon-H1, Building World’s Leading Arabian AI Model (Graphic: AETOSWire)

---