OpenGL is a flexible procedural interface that allows a programmer to describe a variety of 3D rendering tasks. It does not enforce a particular method of describing 3D objects, but instead provides the basic means by which those objects, no matter how described, may be rendered. This mechanistic view of rendering provides for efficient use of graphics hardware, whether that hardware is a simple framebuffer or a graphics subsystem capable of directly manipulating 3D data. OpenGL is rendering-only, so it is independent of the methods by which user input and other window system functions are achieved, making the rendering portions of a graphical program that uses OpenGL platform-independent.
Because OpenGL imposes minimum structure on 3D rendering, it is an excellent base on which to build libraries for handling structured geometric objects, no matter what the particular structures may be. Examples of such libraries include object-oriented graphics toolkits that provide methods to display and manipulate complex objects endowed with a variety of attributes[11][12]. A library that uses OpenGL for its rendering inherits OpenGL's platform independence, making such a library available to a wide programming audience.