We formulate some basic properties of strict approximate solutions, or strict ε-solutions, of convex vector optimization problems. Then saddle-point theorems corresponding to these types of solutions are presented with applications related to duality.