The Atrolux MF02 is rated at a max output of 3000 lm. Given the source, I'd guess it will never actually achieve this. Even if it did, at a modest 100 lm/W, that would be 30W input to the LED. Assuming a rather poor driver efficiency of 80%, that would mean 37.5W load on the batteries. Spread over 4 cells, that's 9.4W per cell. Even given a fully discharged cell (2.7V) that's only 3.5A. Higher driver efficiency, higher LED efficacy, or higher cell voltage would all tend to reduce this value.
So it's very unlikely that under any condition this light will draw more than 3.5A from any cell.
Notice that this calculation does not depend on what the LED configuration (6/12V) or the cell configuration (4S, 4P, 2S2p) are. It is also quite consistent with what Peter Yetman posted. A 2S2P configuration at 3.5A per cell would be 7.0A total, and the measured current at full output is considerably less than that, as anticipated.