Active system ubiquitous to all length scales from a group of bacteria to the human crowd displays collective motion by forming patterns like a vortex, polar flock, cluster, etc. In many practical active systems, the presence of passive - non-motile - particles are observed. We numerically study the collective behavior of a binary mixture of active and passive particles wherein particles are modeled as soft-disks. We used the Lagrangian approach and tracked each particle explicitly at all the time steps. The system dynamics are governed by the interplay of particle-particle interaction, self-propulsion, and alignment interaction. The force equations are solved using the Velocity-Verlet algorithm. Parameters like total energy dissipation, rotational order, and segregation index quantify the system dynamics. We explored the formation of collective milling and phase segregation in the parametric space of coordination parameter, packing fraction, and size dispersity. With an increase in coordination, a randomly distributed initial system undergoes a phase transition to an ordered mill where particles move in a circle around a common center. After attaining the milling phase, the system shows segregation between the active and passive particles. We also observed complete phase segregation by forming a solid passive shell and inner active fluid. Finally, we conclude that a small number of active particles can lead the system to mill formation. Phase segregation can be attained without long-range attraction/repulsion interactions. This work gives a fundamental understanding of phase formation and phase segregation in a binary mixture of the active-passive particles. The findings can be potentially used in cell sorting, segregation of granular particles, pharmaceutical industry, and others.