"So we’ll compare risk at both the 6.2 billion and 21 billion miles traveled levels.
784 cyclists died in 2005 (p. 86). That would make the death rate of 0.37 to 1.26 deaths per 10 million miles.
33,041 motorists/passengers died (p. 86) from 3 trillion miles traveled (p. 15), making their death rate 0.11 per 10 million miles traveled.
So cyclists are either 3.4x or 11.5x as likely to die as motorists, per passenger mile. Neither conclusion is very happy."
Motorcycles are about 30x more dangerous per mile than cars, for reference. I've seen it said that walking somewhere is also more dangerous per mile than driving, but I can't find a source for that.
The page links to a source which suggests cycling is much (6x) safer per hour than driving, which intuitively seems like a more reasonable measure of safety to me, but I'm unsure.