I recently conducted a 10 day experiment on the vampire drain with version 5.8. My data shows at least a 53% improvement between the 5.8 and 4.5 software, and likely more since the 5.8 experiment occurred in a significantly colder garage. On an annual basis, the car will use 474kWh of electricity for vampire drain, enough to drive about 1,250 miles.
For more details on the 4.5 results and the methodology of these experiments, please read the blog post.
In 4.5, the Model S consistently woke up every other day. In 5.8 it woke up either every second or third day. With the 4.5 software, the Model S took up to 45 minutes to recharge to 50% capacity, and with 5.8 it never exceeded 30 minutes. Oddly enough, one incident did not occur at the scheduled 10:00pm time slot, but occurred at 2:00am. I almost missed this data point. The highest usage of kWh by far is the Model S as it pulls about 2.3 kWh in a 15 minute period. The largest other household usage is at most 0.5 kWh in a 15 minute interval, so I am comfortable in assuming that this 2am energy usage was the Model S.
During the 4.5 test, the average outside temperature reported by my utility was 66 degrees Fahrenheit (19 degrees Celsius). During the 5.8 test, California had a very long cold spell and the average temperature was 43 degrees Fahrenheit (6 degrees Celsius). My Model S is parked inside a three car garage facing north. The garage has a very large volume. The ceilings are about 12.5 feet tall (almost 4 meters) and the temperature is pretty consistent and only during very long heat or cold spells is the garage uncomfortably hot or cold. I do not have a temperature record of my garage temperatures but I do not think it is as extreme as the daily averages listed by the utility.
The results for 4.5 showed a Vampire drain of 2.75 kWh per day, or 2,700 miles per year. The results for 5.8 show that number reduced by at least half to 1.3 kWh per day, or 1,260 miles per year. Because of the weather differences, I suspect the true resulting vampire drain in the exact same climate conditions will be better than a 53% improvement. Here is the detailed data from the 5.8 experiment where the car woke up three times. The night of Dec 1, 2013 the car was charged to a 50% reading.
And for completion sake, here is the updated 4.5 chart including the average outside temperature.