The UAE’s Technology Innovation Institute (TII) has unveiled the next generation of its Falcon AI models which it claims outperforms Meta’s most recent Llama 3 model.
Falcon 2 11B and 2 11B VLM were released as open source, making their code accessible to developers globally with unrestricted access.
Falcon 2 11B VLM has visual-to-language capabilities, meaning it can analyse images and generate captions to describe the contents of an image.
The TII stated that it hoped introducing visual-to-language capabilities would create new use cases for AI in industries like healthcare, finance, education and the legal sector.
It stated that both models can run on one graphics processing unit (GPU), touting Falcon’s scalability and deployment ease to businesses wanting to integrate AI into their work.
Both models are multilingual and can work in English, French, German, Spanish and Portuguese.
How well do you really know your competitors?
Access the most comprehensive Company Profiles on the market, powered by GlobalData. Save hours of research. Gain competitive edge.
Thank you!
Your download email will arrive shortly
Not ready to buy yet? Download a free sample
We are confident about the unique quality of our Company Profiles. However, we want you to make the most beneficial decision for your business, so we offer a free sample that you can download by submitting the below form
By GlobalDataExecutive director and acting chief researcher of AI at the TII, Dr. Hakim Hacid, stated that smaller AI models may soon overtake their larger counterparts.
“AI is continually evolving, and developers are recognizing the myriad benefits of smaller, more efficient models,” Hacid stated.
“In addition to reducing computing power requirements and meeting sustainability criteria, these models offer enhanced flexibility, seamlessly integrating into edge AI infrastructure, the next emerging megatrend,” he said.
“Furthermore, the vision-to-language capabilities of Falcon 2 open new horizons for accessibility in AI, empowering users with transformative image to text interactions,” concluded Hacid.
Looking ahead, the TII stated that it would be updating both Falcon models to have a Mixture of Experts (MoE) capability.
MoE will aim to create a more accurate and reliable Falcon that is able to answer industry-specific and niche prompts.
Not only would this help accuracy, but the TII stated that it hoped to cement Falcon’s place as a helper to businesses’ strategic decisions.
Research and analysis company GlobalData forecasts that the total AI market will be worth over $1,037bn by 2030.