I disagree

. The XM-L data sheet specifies luminous flux values at 25C, while the XM-L2 is specified at 85C. That's a big difference.

The cool white U2 bin XM-L has minimum 300 lm at 700mA, calculated 742 lm at 2000mA, 25C.

The cool white U2 bin XM-L2 has minimum 300 lm at 700 mA, calculated 728 lm at 2000mA, 85C.

To compare apples to apples though, you have to use the XM-L2's calculated output at 700mA, 25C. That's 340 lm.

So the XM-L2 is actually substantially brighter UNDER THE SAME CONDITIONS, compared to the XM-L.

As for thermal reasons, they are minimal. The thermal resistance of the package is identical (2.5C/W). The forward voltage spec is a tiny bit lower, but that's probably entirely due to it being specified at 85C instead of 25C.

You do get more light at the same power input, so that means a little less heat. We don't have numbers to calculate accurately, but we can do a little better than a WAG. If the XM-L is 25% efficient at 10W, which seems like a reasonable estimate given numbers I've seen here and there, then the XM-L2 might be 340/300 times 25%, or 28% efficient. So the heat generated would be 7.2W instead of 7.5W. Keep in mind I put together numbers that don't belong together in order to get this estimate, so it's not to be counted on. However, it should be in the ballpark.