In a lot of other models, Arabic/اللغة العربية (and other non-English languages) are more expensive to use because the tokenizer basically punishes other languages by requiring a lot more tokens to represent their text.
Command R improves the tokenization, requiring less tokens to represent these text in these languages, leading to cost savings (on top of the model's raw performance in these languages).
One subtlety worth mentioning is how significant the tokenizer is to the cost to use models in non-english languages. Our tokenizer is meaningfully better than others at the 9 non-English languages, achieving up to a 2x effective cost reduction to use.