The TII-UAE Technology Innovation Institute has just launched Falcon 3: a family of open source AI models with 30 new model checkpoints from 1B to 10B

Advances in large language models (LLM) have created opportunities across industries, from automating content creation to improving scientific research. However, significant challenges remain. High-performance models are often proprietary, which restricts transparency and access for researchers and developers. Open source alternatives, while promising, often struggle to balance computational efficiency and performance at scale. Additionally, limited linguistic diversity in many models reduces their broader usability. These obstacles highlight the need for open, efficient, and versatile LLMs capable of performing well in a variety of applications without excessive costs.

<h3 class="wp-block-heading" id="h-technology-innovation-institute-uae-just-released-falcon-3″>The UAE Institute of technology Innovation has just launched Falcon 3

The technology Innovation Institute (TII) of the United Arab Emirates has addressed these challenges with the launch of Falcon 3the latest version of their open source LLM series. Falcon 3 presents 30 model checkpoints ranging from 1B to 10B parameters. These include Basic models and adjusted to the instructions.as well as quantified versions such as GPTQ-Int4, GPTQ-Int8, AWQ and an innovative 1.58-bit variant for efficiency. A notable addition is the inclusion of Mamba-based modelsthat leverage state space models (SSM) to improve inference speed and performance.

By launching Falcon 3 under the TII Falcon-LLM 2.0 LicenseTII continues to support open commercial use, ensuring broad accessibility for developers and businesses. The models are also compatible with the flame architecturemaking it easy for developers to integrate Falcon 3 into existing workflows without additional expense.

Technical details and key benefits

Falcon 3 models are trained on a large-scale data set of 14 billion chipsa significant jump from previous iterations. This extensive training improves the models' ability to generalize and perform tasks consistently. Falcon 3 supports a 32K context length (8K for the 1B variant), allowing it to handle longer entries efficiently, a crucial benefit for tasks such as summarizing, document processing, and chat-based applications.

The models retain a Transformer-based architecture with 40 decoder blocks and employ Grouped Query Attention (GQA) presenting 12 query heads. These design choices optimize computational efficiency and reduce latency during inference without sacrificing accuracy. The introduction of 1.58-bit quantized versions allows models to run on devices with limited hardware resources, offering a practical solution for cost-sensitive deployments.

Falcon 3 also addresses the need for multilingual capabilities by supporting four languages: English, French, Spanish and Portuguese. This improvement ensures that models are more inclusive and versatile, serving diverse global audiences.

Results and insights

The Falcon 3 benchmarks reflect its strong performance across all evaluation data sets:

• 83.1% in GSM8K, which measures mathematical reasoning and problem-solving abilities.
• 78% at IFEval, showing its instruction following capabilities.
• 71.6% at MMLU, highlighting strong general knowledge and understanding across domains.

These results demonstrate Falcon 3's competitiveness with other leading LLMs, while its open availability sets it apart. Expanding the parameters from 7B to 10B has further optimized performance, particularly for tasks that require multitasking reasoning and understanding. Quantized versions offer similar capabilities while reducing memory requirements, making them suitable for deployment in resource-constrained environments.

Falcon 3 is available in hugging faceallowing developers and researchers to experiment, tune, and deploy models with ease. Support for formats such as GGUF and GPTQ ensures seamless integration into existing toolchains and workflows.

Conclusion

Falcon 3 represents an important step forward in addressing the limitations of open source LLMs. With its range of 30 model checkpoints, including basic, instruction-tuned, quantized and Mamba-based variants, Falcon 3 offers flexibility for a variety of use cases. The model's strong performance across all benchmarks, combined with its efficiency and multilingual capabilities, makes it a valuable resource for developers and researchers.

By prioritizing accessibility and commercial usability, the UAE technology Innovation Institute has solidified Falcon 3's role as a practical, high-performance LLM for real-world applications. As ai adoption continues to expand, Falcon 3 is a clear example of how open, efficient and inclusive models can drive innovation and create broader opportunities across industries.

Verify he Models hugging faces and Details. All credit for this research goes to the researchers of this project. Also, don't forget to follow us on <a target="_blank" href="https://twitter.com/Marktechpost”>twitter and join our Telegram channel and LinkedIn Grabove. Don't forget to join our SubReddit over 60,000 ml.

Trending: LG ai Research launches EXAONE 3.5 – three frontier-level bilingual open-source ai models that deliver unmatched instruction following and broad context understanding for global leadership in generative ai excellence….