If we keep selecting AI models one after another they keep getting added to cache. Expected behaviour is that the old model is taken out of cache before adding the new model of cache.