GPT-4 vs. GPT-3.5 AS CODING ASSISTANTS
Author:
Moussiades Lefteris1, Zografos George1, Papakostas George1
Affiliation:
1. International Hellenic University
Abstract
Abstract
Large Language Models are not limited to understanding and producing natural language. Instead, they do equally well in understanding and generating source code in various programming languages. At the time of writing, GPT-4 was considered the most potent large language model from OpenAI. This paper compares GPT-4 with its immediate ancestor, GPT-3.5, as coding assistants. More specifically, we have constructed appropriate tests to check whether the two systems can a) answer typical questions that can arise during code development, b) generate reliable code in response to user requirements, and c) contribute to code debugging. The test results for both models are impressive. However, the performance of GPT-4 is outstanding and signals an increase in the productivity of programmers and the reorganization of software development procedures based on these new tools.
Publisher
Research Square Platform LLC
Reference24 articles.
1. Austin, J., Odena, A., Nye, M., Bosma, M., Michalewski, H., Dohan, D., Jiang, E., Cai, C., Terry, M., Le, Q., Sutton, C.: Program Synthesis with Large Language Models. (2021). https://doi.org/10.48550/ARXIV.2108.07732 2. BigDecimal (Java Platform SE 8). (n.d.). Retrieved January 24, from (2024). https://docs.oracle.com/javase/8/docs/api/java/math/BigDecimal.html 3. Programming with ChatGPT: How far can we go?;Bucaioni A;Mach. Learn. Appl.,2024 4. Chen, M., Tworek, J., Jun, H., Yuan, Q., Pinto, H.P., de Kaplan, O., Edwards, J., Burda, H., Joseph, Y., Brockman, N., Ray, G., Puri, A., Krueger, R., Petrov, G., Khlaaf, M., Sastry, H., Mishkin, G., Chan, P., Gray, B., Zaremba, S.: W. Evaluating Large Language Models Trained on Code. (2021). https://doi.org/10.48550/ARXIV.2107.03374 5. Christopoulou, F., Lampouras, G., Gritta, M., Zhang, G., Guo, Y., Li, Z., Zhang, Q., Xiao, M., Shen, B., Li, L., Yu, H., Yan, L., Zhou, P., Wang, X., Ma, Y., Iacobacci, I., Wang, Y., Liang, G., Wei, J., Liu, Q.: PanGu-Coder: Program Synthesis with Function-Level Language Modeling. (2022). https://doi.org/10.48550/ARXIV.2207.11280
|
|