it's a tricky thing to give a simple answer to, because not all complications react the same way to blood glucose concentrations, and because most studies have been done on average blood sugar over time rather than on peaks and troughs.
So for example an answer often given is that a reduction of 1% in hba1c (average blood sugar over 3 months) reduces your risk of long-term complications by 20-25%.
But this isn't really that clear. For a start, Type 1 and Type 2 diabetics have different risk factors even at the same level of blood glucose (eg T2s are slightly more likely to get neuropathy, T1s slightly more likely to get retinopathy - I think - don't quote me on that.) Some complications, especially those in the eyes, are much more sensitive to changes in blood glucose than others. So for example if you've been running at a high level and you drop several % of hba1c quickly, you actually increase your risk of eye problems. Some parts of the body can tolerate higher bg levels than others. There are lots of other factors that are linked to diabetes but not to blood sugar - eg there's work being done on eye problems which suggests that some people are just more susceptible to it whatever they do with their blood glucose. And by talking purely about risk of long-term complications, it glosses over the increased risk of potentially lethal severe hypos which some (not all) studies found in insulin users who reduced their Hba1c.
The other obvious problem with the above statement (20-25% thing) is that it depends on your starting point. If your hba1c is 12% and you drop to 11% that's great. But if your hba1c is 6% and you drop to 5%, is it still as big a risk reduction? We don't really know.