The paper evaluates the capabilities of large language models (LLMs) in generating visualizations. It investigates whether LLMs can create a wide variety of charts, use different libraries, and configure visual variables effectively. The study tests ChatGPT3 and ChatGPT4 with libraries such as matplotlib, Plotly, and Altair. The results show that ChatGPT4 can generate almost 80% of the proposed charts. It performs well in generating code for visualizations, with some issues in specific chart types like bullet charts and range plots. The study also explores how LLMs handle different libraries and data types, and finds that the performance is largely consistent across libraries. The paper highlights that while LLMs can generate visualizations, there are still challenges in configuring visual variables and ensuring the accuracy of the generated code. The results suggest that LLMs have significant potential for visualization tasks, but further research is needed to improve their performance and usability. The study provides a comprehensive analysis of the capabilities of LLMs in generating visualizations, and offers insights into the current state of the technology. The findings indicate that LLMs can be a valuable tool for visualization tasks, but there are still limitations that need to be addressed. The paper concludes that while LLMs show promise, they are not yet perfect and require further development to become more reliable and user-friendly. The study also emphasizes the importance of continued research in this area to improve the capabilities of LLMs in generating visualizations.The paper evaluates the capabilities of large language models (LLMs) in generating visualizations. It investigates whether LLMs can create a wide variety of charts, use different libraries, and configure visual variables effectively. The study tests ChatGPT3 and ChatGPT4 with libraries such as matplotlib, Plotly, and Altair. The results show that ChatGPT4 can generate almost 80% of the proposed charts. It performs well in generating code for visualizations, with some issues in specific chart types like bullet charts and range plots. The study also explores how LLMs handle different libraries and data types, and finds that the performance is largely consistent across libraries. The paper highlights that while LLMs can generate visualizations, there are still challenges in configuring visual variables and ensuring the accuracy of the generated code. The results suggest that LLMs have significant potential for visualization tasks, but further research is needed to improve their performance and usability. The study provides a comprehensive analysis of the capabilities of LLMs in generating visualizations, and offers insights into the current state of the technology. The findings indicate that LLMs can be a valuable tool for visualization tasks, but there are still limitations that need to be addressed. The paper concludes that while LLMs show promise, they are not yet perfect and require further development to become more reliable and user-friendly. The study also emphasizes the importance of continued research in this area to improve the capabilities of LLMs in generating visualizations.