Selecting the appropriate texts as the authentic material for English teaching, particularly at the university level, matched with students' mastery level is still challenging. This study attempts to investigate the readability level of reading texts through the framework of Systemic Functional Linguistics (SFL) issued by ChatGPT, focused on whether the complexity level of the texts matches the standard level of the Common European Framework of Reference (CEFR) by proposed British Council. Through the qualitative design with content analysis, this study examines eighteen texts classified following CEFR levels issued by ChatGPT. It analyzes their lexical density, grammatical intricacy, and lexical variation to know the complexity of the texts. The analysis revealed that the complexity level of the texts issued by ChatGPT has not thoroughly followed the standard level of the texts used by the British Council on the CEFR level, indicating that several texts are inappropriate with their complexity levels. It is found that the complexity level of the texts issued by ChatGPT is determined by the length of the texts that increase from the basic level to the proficient level seen from the number of words in each text, not by their lexical density, grammatical intricacy, and lexical variation indexes. The study provides implications for the lecturer to carefully select the text used as teaching material issued by ChatGPT by relying on the lexico-grammatical analysis of the text to help students achieve the comprehension stage.