I didn´t get the message "jitter couldn´t find destination named ml" after clicking in "name ml".
You have another conflict: both "normal" jitter video and the OpenGL model are rendering in the same context/window. Sometimes this means you get a flickering image, because the program is trying to display the 2 images at the same time.
I see that you want to mix the 4 videos using jit.op, in normal CPU mode: this means that you have to bring back the model from the OpenGl realm to the matrices realm. You can do this by using a matrix as a rendering context (a matrix named "ml"). See the jitter example "jit.gl.render-tomatrix", there are several methods to do this.
jit.gl.handle does not work in the previous example, because the output window does not belong to OpenGL, so it does not receive the movement of the mouse.
I found this solution, rendering the model in a second window where you can move it and rotate it. jit.gl.asyncread will then transfer the (manipulated) model to normal CPU matrices. Maybe there are smarter solutions, but this works fine
sorry, I am a little bit confused now as your patch is behaving as expected. You see a big gray screen when you load your cubo.obj because of its size. Add the next attribute to your model object: [poly_mode $1 $1] and then a toggle to it, that enables/disables the wireframe rendering in openGL. When on, you will see that the cube displays the vertices that define the object's shape. Another thing that you can try is to use the scale attribute. Check the jit.gl.model help file.
About the last question, sorry, i don't seem to understand. If you want to bind textures to your object you can either do it inside jitter or with your modeling software(the textures will be created on a separate file). Be aware that jitter doesn't load all types of bind materials and textures. I remember(but i am not sure anymore) C4D had some issues with exporting the textures. I think blender might be a better option for such task(you can even use shaders , very cool!):