I was working on my university studies, particularly on some LLM projects on the Arabic language, such as hULMona. It was a challenging task due to the diverse nature of the Language and the extensive fine-tuning required. However, with the advent of the Generative models era, this process has become far more efficient, and LLMs are continually enhancing their capabilities. Even though, we still encounter significant challenges when it comes to the Arabic language. Despite the advancements, the best generative models available still fall short of speaking or writing Arabic accurately. Nonetheless, I was pleasantly surprised to discover an open-source model on YouTube that seemed to yield promising results, even for the Arabic language. I have included below, images of some prompts that I tested, to compare between the most powerful model by OpenAI which is ChatGpt4o and the Cohere Command R+, and the results speak for themselves as you'll see on the images. For those seeking more information, feel free to explore the Cohere platform and its documentation website: "docs.cohere.com". Additionally, the Cohere-toolkit is available on GitHub for further insights.

Thoes are the list of prompts I have tested out on brand new chat of the two models without any initial prompts :