As you may know, a definition of phase that you can find in Callister is "any homogeneous portion of a system/microstructure having distinguishable physical and/or chemical properties".
Although GP zones are clusters of solute, and thus may show locally different chemical properties than the matrix, they are coherent with the matrix (they form so that there is a definite relationship between the crystal structure of the matrix and the GP zone; both are FCC and there is usually univocal correspondence between atoms at each side of the boundary between the matrix and the GP zone) and from my experience they are not usually termed "phases" or "precipitates" in the scientific literature. They are usually regarded as clusters of solute and that's all.
According to the diffinition of the phase: The macroscopic homogeneous part of inhomogeneous system separated by the interface from other part of the system. According to this definition it is not a phase (no interface) and besides, we can not say that it is macrosocpic part because consist only certain and small number of atoms of each types.
I agree with Dr. Jose and Dr. Rodin that GP zone can not be strictly treated as a phase if one goes by the definition of a phase.
But I need clarification on the treatment of GP zones as a new phase precipitate based on the computer simulation of spinodal decomposition and formation of GP zones (GP zones is generally understood as the cluster of solute where the periodicity is destroyed, and X-rays are sensitive to such lack of periodicity).
GP zone formation by spinodal decomposition:- I quote from the book by Armen G. Khachaturyan, Theory of Structural Transformations in Solids, 2008 (Dover edition) , page 151:- "one of the reasons a GP zone is regarded as different from the usual new phase precipitate is its extreme thickness, of the order of 1 to 3 crystal lattice planes. This was considered incongruous with conventional thermodynamics. The computer simulation of the process, which used conventional thermodynamics, however, shows that such microscopically thin precipitates can be formed and that clearly GP zones are nothing other than precipitates (see Fig. 50c)".
Please see (1) paper by K.T. Kashyap and Praveennath Koppad in RG:- https://www.researchgate.net/publication/232734563_Small-angle_scattering_from_GP_zones_in_Al-Cu_alloy
(2) Wikipedia:- "Personal Reminiscences" by A. Guinier (please see link if you have access:-