ModelGPT: Cache builder responses

Store responses from ModelGPT locally, significantly accelerating the model generation time. Plus, you can now effortlessly store and reutilize your preferred responses whenever required in the future.