Most companies prepare just a single budget scenario, which is their best guess regarding how the next year will turn out. This scenario is based upon a range of supporting assumptions, any one of which can lead to diverging results - and usually does. So, though you may spend a considerable amount of time on that "mainstream" budget scenario, just that one version will not be enough to prepare you for what may - and probably will - happen.
It makes sense to add two more scenarios, one for the absolute worst case, where bankruptcy is looming, and one for the most phenomenal sales success. Sounds unlikely that either one will ever happen? If you don't plan for success, it never will happen, and bankruptcy scenarios are far more frequent than you might think. Consequently, it is useful to know what resources you'll need for a phenomenally successful year, and how deep you will have to cut to avoid bankruptcy. Is that enough scenarios? No.
There are gaping holes between the two opposite-extreme scenarios and the mainstream version. Realistically, actual results will fall into either of those two holes, so you should spend some time figuring out what you will do for situations that are somewhat above and below the mainstream scenario.
So the answer is - five budget scenarios. However, if some of your underlying assumptions are more likely than not to occur or to fail, then you may want to drum up some extra models just for those specific situations.
All of this talk of multiple models does not mean that you should spend an equal amount of time on each one. The mainstream scenario requires the most work, because it is (presumably) the most likely, with less work needed for the less likely ones. Nonetheless, you should at least spend time determining financial results at a high level for each scenario, and conceptualize what those situations will do to the company's operations.