The main goal of the Astrometric Gravitation Probe mission is the verification of General Relativity and competing gravitation theories by precise astrometric determination of light deflection, and of orbital parameters of selected Solar System objects. The key element is the coherent combination of a set of 92 circular entrance apertures, each feeding an elementary inverted occulter similar to the one developed for Solar Orbiter/METIS.1 This provides coronagraphic functions over a relevant field of view, in which all stars are observed for astrometric purposes with the full resolution of a 1 m diameter telescope. The telescope primary mirror acts as a beam combiner, feeding the 92 pupils, through the internal optics, toward a single focal plane. The primary mirror is characterized by 92 output apertures, sized according to the entrance pupil and telescope geometry, in order to dump the solar disk light beyond the instrument. The astronomical objects are much fainter than the solar disk, which is angularly close to the inner field of view of the telescope. The stray light as generated by the diffraction of the solar disk at the edges of the 92 apertures defines the limiting magnitude of observable stars. In particular, the stray light due to the diffraction from the pupil apertures is scattered by the telescope optics and follows the same optical path of the astronomical objects; it is a contribution that cannot be eliminated and must therefore be carefully evaluated. This paper describes the preliminary evaluation of this stray light contribution.