The computer is getting more powerful every day and creating animations for a 3D model ”on the fly” rather than pre-made animations is getting more feasible. A 3D model that is composed by segments and joints, formed for example like a human, can also be viewed as a kinematic model, the joints can be modified such that different postures can be achieved for the model. A kinematic model have a hierarchic tree structure with a parent-child relationship. A change made to a parent propagates to its children. Interacting with a kinematic model can be done in two distinct ways, forward kinematics and inverse kinematics. The most common way of interacting with a kinematic model is by using forward kinematics, which works by changing each joint until a desirable posture is achieved. The other way, known as inverse kinematics is more complex and needs more computer resources, only need a direction or a position for any joint of where it should be. The inverse kinematic methods are all using a matrix called Jacobian matrix, which consists of all first-order partial derivates derived from a kinematic model. The main goal of this thesis is to evaluate current methods solving Inverse Kinematics and to see if it possible to create animations ”on the fly” for a model.