Generative Artificial Intelligence and Large Language Models (LLMs) are proven effective in various applications, including problem formulation and knowledge extraction. This study evaluates the capability of LLMs to generate functional models of Technical Systems, including identifying both useful and harmful functions. The objective is to assess whether modern LLMs can facilitate basic functional analysis based on user-provided inputs. A set of ten dummy problems was designed to test LLM-based functional modeling. Functional models generated by LLMs are compared against human-based analysis, with evaluation criteria focusing on accuracy in function identification and parameter definition. The findings indicate that LLMs can be used to extract proper TRIZ functions from input text. LLMs demonstrate promise in supporting fundamental functional modeling tasks, particularly in synthesizing and structuring knowledge for early-stage design analysis. Further research is required to refine these models for comprehensive functional analysis and domain-specific applications.