circle packing 26d code release

The evaluator for the 26D circle packing problem can potentially be hacked because the required accuracy is only 10e-6. Could you release the results for this problem? I’m concerned the solution may no longer be correct if the accuracy threshold is tightened to 10e-9 or smaller.