Here's how I'd approach it. Nominal is just the label on the bottle, not exactly what's in it. Your measurements are "actual", not nominal (IMO if I'm understanding your post correctly). As an example, pull the gas specs from the local gas supplier. For any given mixed gas, they'll spec the actual composition of any given bottle to be the labeled nominal composition +/- some percent. For a bottle of 50% Ar - 50% He, it might be +/-5 %, or +/-2% or whatever, so while the nominal is 50/50, the actuals might be 55/54, 52/48 or whatever. To give you an idea of what sort of "universe" of tolerance you're working in, find a comparable commercially available gas mix and see what it's tolerance is. If the tolerance of your mix is within what they specify for theirs, then you have some good evidence for your non-conformance report to support the welds that have been produced are good. If the tolerance of your mix is outside what industry is using for a comparable gas, then you probably have some decisions to make. Basically, since there is little guidance with in the code for this, I'd use the industry accepted nearest standard as the yardstick. And then once this is all sorted out, it'd be a wise idea to write it down and roll it into your quality program/procedures.
from a technical standpoint, depends what the major and minor gas is. I'm guessing 84 He, 16 Ar? below 25% gas conductivity (penetration, bead shape) is very directly correlated to composition so small changes can make differences. easiest way is to make a few test welds cut them up and look at the macros to see what the penetration or heat affected zone looks like. In all reality if it's manual welding it probably wont' make a big enough difference.