The difference should exist, but in routine PCR work, the dNTPs are at high concentrations that diminish the difference. So the concentrations of dNTPs matter. It may be predicted that at a critically low concentration of dNTPs, PCR for a sequence with a higher GC content will occur more readily than that with a lower GC content.
the GC content of your primers has no bearing what so ever on PCR efficiency unless high average or local content causes or predisposes to primer dimers; linear intra primer annealing or intra primer annealing via hairpin loops; In contrast the GC content of your target does have an effect on PCR efficiency if that content is over 70% - causing a reduction in PCR/primer efficiency and/or you have regions with high local GC content which can cause the polymerase to stall and result in low amplification efficiency, especially for PCR amplicons over 1kb; Specifically
1. High local GC content in primers - especially if your last 5bp on either primer contain 4 GC residues - can reduce PCR efficiency by causing primer dimer formation
2. Screen for primer dimers by using programs like IDT Oligo analyser
https://www.idtdna.com/calc/analyzer
3. Screen for hair pin loops using Oligo calculator
Even if your target has a low GC content ( < 70%) or no regions with more than 5 GC residues intra or inter primer annealing can reduce PCR efficiency. If identified make alternative primers for this reason
4. In contrast if no such intra or inter primer annealing is discovered by the above (or similar) programs but your target sequence has high GC content or average GC content but contains regions with 5 or more successive GC residues the former can lead to secondary structure which reduces PCR efficiency and the latter can cause the polymerase to stall and detach; also reducing PCR efficiency
5. To investigate whether the GC content and distribution within your target sequence predisposed to secondary structure and therefore might reduce PCR efficiency check with M Fold
http://unafold.rna.albany.edu/?q=mfold
6. Note high local GC content or average PCR content will reduce PCR efficiency particularly if the target sequence is > 1kb
7. Preferably choose an alternative target sequence but if this sequence has to be used add 5% DMSO and/or 0.5 M betaine to your PCR reaction
8. In addition preceded your main PCR by 1 denaruting cycle = 5 min at 95c
9. In addition for targets with high GC content use primers longer than 21bp and/or high GC content with a high Tm and perform both primer annealing and extension at the same temperature (above 65C). This 2 step PCR can assist with PCR efficiency for targets with high GC content as well
I was comparing the efficiency between GC-rich and AT-rich situations (if that's what the original question is about). I thought the question is about how GC-rich (whether in the primer or the template) compares with AT-rich in general, not about whether GC content per se has bearing on PCR reactions. Problems associated with local sequences (e.g. hairpin) can also occur with AT-rich scenarios.
I take your very important point about AT rich sequences - both of primer and target sequence - affecting PCR efficiency; In fact, apart from situations outlined which lead to primer dimers AT rich motifs particularly in primers can have a profound impact on PCR efficiency
To add therefore to Mings comments and effectively finish my dialectic on PCr efficiency, these are the principal issues with high AT rich sequences/very low GC rich sequences associated with primers and target sequence:
1. Primers that are AT rich can lead to poor PCR efficiency; Specifically, primers that end in AT/ AA and particularly TT can culminate in poor priming and inefficient PCR
2. Also, in the last 5bp of your primer sequences try not to incorporate 3 or more A/T residues
3. For most efficient priming terminate your primers with GG; GC; or CC. Avoid 3 or more G/C residues in the last 5bp
4. With reference to target sequence, high AT and low GC can lead to polymerase stalling and detachment culminating in low PCR efficiency. Thus, where possible, avoid AT replete target sequences