We have completed a survey and inter-comparison of several 10 Be and 26 Al standard reference materials (SRMs) that are in routine use at various AMS laboratories to assess their relative values and the accuracy of their quoted nominal ratios. The accelerator measurement cycle, analysis procedure and setup used at the ANTARES AMS facility for this survey are described. We focused on a new set of 10 Be and 26 Al serial dilutions of standard reference materials (SRMs) prepared by Kuni Nishiizumii at the University of California, Berkeley, and found excellent systematic reproducibility and internal consistency. For other standard materials, minor deviations are evident even when the results have been recalibrated to a common half-life. In particular, we confirm that the NIST 10 Be SRM-4325 has a 14% greater 10 Be/Be ratio than that certified by NIST when it is calibrated against other SRMs whose ratios have been normalized to a common 1.5 Ma 10 Be half-life. In order to investigate this apparent discrepancy, we report on the results of an absolute, normalization independent, measure of the NIST-4325 10 Be/Be ratio. Within the constraints of this type of measurement and its systematic errors, we determine an absolute value for the 10 Be/Be SRM-4325 ratio in the range 26,050 to 24,800 โข 10 ร15 in support of the certified value of 26,800 โข 10 ร15 given by NIST. We hesitate to directly infer as a consequence that the 10 Be half-life is 1.34 Ma because such an inference is contingent on a direct and accurate specific activity in the parent solution, which at present is not available.