It's 23F here - coldest in 3 years I believe, so I measured a couple things
I used a Burton 8 ounce canister which is supposed to be Butane and Propane but I'm not sure. $2.50 for 8 ounces at Fred Meyer. Maybe it's just whatever happened to come out of the well.
I got a SOTO - the new Windmaster. There's a lot of talk about how SOTOs with their regulator work better at cold temperature, but no data just operating at specific temperature. Just video of filling up container with ice water or whatever.
So I heated up 15.6 oz of water with SOTO and Pocket Rocket - standard needle valve stove. Measured weight before and after to calculate fuel use. Measured temp before and after. I tried to get it to 90C but measured actual and normalized the data.
Test setup - cooking thermometer through hole in top:
SOTO - 9.2 to 93.3 C, 6.2 g fuel - 12.8 g/L/80C
Pocket Rocket - 11.1 to 93.1 C, 5.7 g fuel - 12.1 g/L/80C
They both took 6 minutes. It normally takes 3 minutes and uses about the same amount of fuel.
So, the SOTO and Pocket Rocket slow down by the same amount - SOTO offers no advantage at cold temperature. I think I'm getting down towards the lower temperature limit for this canister.
The fuel usage was a little less for Pocket Rocket, but that could easily be measurement error.
Of course, there's advantage to not having to adjust needle valve...
Then, I tried the trick of wrapping cooper wire around canister and the end into the flame. #16 wire I believe. 20 inches long. 0.33 ounces:
SOTO with wire - 10.3 to 94.7 C, 6.7 g of fuel - 13.8 g/L/80C
It took 3 minutes and 10 seconds. This is the same as at warm temperatures.
It did take a little more fuel. Could be measurement error but this is less likely because difference is greater. Probably it's because the flame was higher which makes it less efficient - more heat blows away into the air.
So, conclusion is for 0.33 ounces and some fiddling, you can operate at maybe 10 degree F colder.