The Varian gradient pump design is somewhat unique: it has multiple inlet valves in one high-pressure pump head, one valve per solvent. This is not what most people mean by "high-pressure" mixing. Mixing still takes place at low pressure, but there is no separate proportioning valve. Mixing takes place inside the pump head during the fill stroke.
The Agilent uses more the classic design: a quadruple proportioning valve assembly mixes the solvents before they enter the high-pressure pump head. Dionex uses this design too, but I think we did it a little better.
Both designs can deliver excellent results.
Most people use the term "high-pressure mixing" to refer to a different design: a separate pump for each solvent, and the mixer is after the pumps. The pumps change flow rates to make the gradient. The mixing volume for this design can be very small. Agilent (and Dionex) offer this as a more expensive option.
What you want to look at is the gradient delay volume. This takes into account all the various volumes between the point of mixing and the top of the column. (Of course for an isocratic method this does not matter much.) For standard 1 mL/min gradient applications, it usually makes little difference. If you go to 0.2 mL/min then the difference becomes much more important.
