I'll ad a option 2.1: They were full of shit, but it is because things are ridiculously chaotic and trying to predict it would be folly so they just went "lol 10 million".
They seem to be relying on people not understanding inflation rates with their 10 million number, too.
For example, if the inflation rate is 1000%, then prices are 10 times higher, so you could say "at this rate, next year there will be 10000% inflation!" ... but that's not how it works. But there's no other way that I can put the IMF's 10 million number into context.
They don't seem to be taking any actual figures into account. We have the exchange rate for example (dolartoday.com), we can see how many USDs you get for a Venezuelan Bolivar, and we can look at Bloomberg's data on the change in the price of a cup of coffee in Caracas and see that that aligns perfectly with dolartoday.com's rate.
Empirical evidence like the price of a cup of coffee is in fact good data. If the real inflation rate was IMF-high they'd have to charge a lot more for that cup of coffee. It could be fun to predict what the cup of coffee should cost by the end of next year according to the IMF. It's 450 bolivars now. If the inflation for 2019 really was 10 million percent, then the cup of coffee should cost 45 million Bolivars by the end of 2019. That's the IMF prediction as of now.
However, just found a recent article, and Stephen Henke from Cato actually calculated the real inflation rate at the end of the year as 80,000%.
https://www.forbes.com/sites/stevehanke/2019/01/01/venezuelas-hyperinflation-hits-80000-per-year-in-2018/80000% for 2018, not the IMF's 1.37 million%. Which leaves the IMF's error rate at ~95% off. Also note that the overall rate since August was 170,000%, which is twice as high as Henke's ending figure for the year, which suggests the rate itself is declining, not increasing. It's reasonable to assume a trend where the rate halves every 6 months, to 20,000% for 2019. That would leave the IMF's error rate for 2019 as 99.98% incorrect.
For future reference, the IMF's 2019 10-million% prediction would mean the cup of coffee costs 45 million bolivars by this time next year. My 20,000% prediction would put the cup of coffee at about 90,000 Bolivars. Sure, that's a huge increase, but you can see why the IMF figure is in another ballpark.
Assuming the rate neither goes up or down from Henke's 2018 figure would put the coffee at 360,000 Bolivars. I'll come back and assess the coffee situation in about 6 months or so.