I don't know if my idea will come through in this post, but I will give it a try. Maybe what I say will cause you to find additional solutions.
First, I would probably still create a gradebook category that will contain all of these quizzes, just in case you eventually want to add other grade categories, but this is probably optional for you. I would use "simple weighted means of grades" for this category.
Don't mess with anything when you're creating the quiz. Leave it at one point per question, and set the total for each quiz at 10.
In the gradebook, when you are staring at "categories and items", to the right, you will see that you have the ability to "edit" the quiz. If you click on this, you will see that you can use a multiplier and and offet. This is where I would give bonus for level of difficulty. Using the offset and multiplier is a little tricky, so make sure to experiment and get the hang of it. For example, an orage quiz can have a multiplier of .5, etc.
When the student submits the a quiz (I assume these will be automatically graded), they see their score out of 10. But when they go to the gradebook, the gradebook will show their weighted grade.
Initially, it will take you a little effort to "edit" each color quiz. But once you have it set up, you won't have to mess with it for each time the course is re-offered.
What might be the trickiest thing to do is to decide what you want the student to see in their gradebook, and how you really want the weights to affect their overall grade. But create a dummy student, log in as this student, take a quiz, then go to the gradebook to see if your logic is correct.
By the way, if you want their actual score to always show in the gradebook, then you would have to consider using an "equation" to calculate their final grade. This too can be done in Moodle.
One more "by the way", if you are coming to Moodle from a product like Blackboard, Moodle's gradebook is much more powerful, it makes Blackboard look like a wimp. The gradebook is also more powerful than Desire2Learn.
For whatever reason, I haven't found the need to use weighted means, but there is probably an approach to your problem with it.
Hope this helps.