General relativity considers gravity as the curvature of space-time instead of a force. Therefore, what is bent around an astronomical object is the space-time itself. The light just follows the path as regular.
In classical physics however, I think we would interpret gravitational lensing as the light's being pulled by a massive object (correct me if I am wrong). This implies some centripetal acceleration on light.
The question is does the frequency of light have any effect on the amount of bending in gravitational lensing? Is it possible to separate the light into its components by gravity, like a glass prism would do?