Frontier AI Model Output for Educational Application Development via Vibe-Coding: A Systematic Literature Review Comparing Western and Eastern AI Models

Contemporary artificial intelligence models can translate natural-language prompts into executable code, a practice that practitioners have termed vibe-coding. The capability is offered by Western developers (OpenAI, Anthropic, Google, Microsoft) and Eastern developers (DeepSeek, Alibaba, Zhipu AI, Moonshot AI) alike, yet no systematic comparison of their output quality in educational settings has been published. This review addresses that gap. Drawing on the Technological Pedagogical Content Knowledge framework and Rogers' Diffusion of Innovation theory, a PRISMA 2020 systematic review was conducted on 42 studies published between January 2022 and August 2026 and retrieved from Scopus, Web of Science, IEEE Xplore and the ACM Digital Library. Each study was assessed against six dimensions: code correctness, code quality, hallucination rates, natural language understanding, developer productivity and pedagogical appropriateness. The evidence indicates convergence rather than supremacy. Eastern models perform within the range established by Western models across several coding benchmarks, and independent testing places DeepSeek V3.2 inside the hallucination band occupied by four Western frontier models. The supporting evidence, however, remains markedly asymmetrical. Every one of the seven Eastern-model entries in the corpus is a technical report published by the organisation that built the model, and this holds for the 2026 frontier releases as firmly as for their predecessors. One 2026 flagship release was excluded outright because its technical report states no measurable capability outcome at all. Western models have been examined through classroom deployments, independent audits and large-sample quasi-experimental designs, whereas Eastern models are evaluated predominantly through technical reports issued by their own developers. Work published in 2026 has begun to address vibe-coding security and cross-model hallucination, but no head-to-head classroom comparison of Western and Eastern models has yet been reported. That absence is identified as the principal gap in the field. The review proposes a reusable six-dimension framework for future comparative research and sets out the implications for educators and policymakers operating within an increasingly bipolar artificial intelligence landscape in Southeast Asia and beyond.

Keywords

systematic literature review, vibe-coding, AI-assisted code generation, frontier AI models, code output quality

Downloads

PDF JATS XML

References

1. Almanasra, S., Suwais, K., & AlRamouni, S. (2026). A comparative evaluation of ChatGPT-4o and DeepSeek-R1 for AI-generated code quality using software engineering metrics. Array, 30, Article 100979. https://doi.org/10.1016/j.array.2026.100979 [Google Scholar] [Crossref]

2. Becker, B. A., Denny, P., Finnie-Ansley, J., Luxton-Reilly, A., Prather, J., & Santos, E. A. (2023). Programming is hard, or at least it used to be: Educational opportunities and challenges of AI code generation. In Proceedings of the 54th ACM Technical Symposium on Computer Science Education V. 1 (pp. 500–506). Association for Computing Machinery. https://doi.org/10.1145/3545945.3569759 [Google Scholar] [Crossref]

3. Braun, V., & Clarke, V. (2006). Using thematic analysis in psychology. Qualitative Research in Psychology, 3(2), 77–101. https://doi.org/10.1191/1478088706qp063oa [Google Scholar] [Crossref]

4. Campbell, M., McKenzie, J. E., Sowden, A., Katikireddi, S. V., Brennan, S. E., Ellis, S., Hartmann-Boyce, J., Ryan, R., Shepperd, S., Thomas, J., Welch, V., & Thomson, H. (2020). Synthesis without meta-analysis (SWiM) in systematic reviews: Reporting guideline. BMJ, 368, l6890. https://doi.org/10.1136/bmj.l6890 [Google Scholar] [Crossref]

5. Cao, R., Chen, M., Chen, J., Cui, Z., Feng, Y., Hui, B., Jing, Y., Li, K., Li, M., Lin, J., Ma, Z., Shum, K., Wang, X., Wei, J., Yang, J., Zhang, J., Zhang, L., Zhang, Z., Zhao, W., & Zhou, F. (2026). Qwen3-Coder-Next technical report [Preprint]. arXiv. https://doi.org/10.48550/arXiv.2603.00729 [Google Scholar] [Crossref]

6. Chen, A. J., Cao, Y., Shao, M., Karri, R., & Shafique, M. (2026). Code for all: Educational applications of the “vibe coding” hackathon in programming education across all skill levels [Preprint]. arXiv. https://doi.org/10.48550/arXiv.2604.22747 [Google Scholar] [Crossref]

7. Churilov, A. (2026). The range shrinks, the threat remains: Re-evaluating LLM package hallucinations on the 2026 frontier-model cohort [Preprint]. arXiv. https://doi.org/10.48550/arXiv.2605.17062 [Google Scholar] [Crossref]

8. DeepSeek AI. (2024a). DeepSeek-Coder-V2: Breaking the barrier of closed-source models in code intelligence [Preprint]. arXiv. https://doi.org/10.48550/arXiv.2406.11931 [Google Scholar] [Crossref]

9. DeepSeek AI. (2024b). DeepSeek-V3 technical report [Preprint]. arXiv. https://doi.org/10.48550/arXiv.2412.19437 [Google Scholar] [Crossref]

10. Deng, J., Fan, Z., & Meng, R. (2026). Understanding the (in)security of vibe-coded applications [Preprint]. arXiv. https://doi.org/10.48550/arXiv.2606.23130 [Google Scholar] [Crossref]

11. Deng, R., Jiang, M., Yu, X., Lu, Y., & Liu, S. (2024). Does ChatGPT enhance student learning? A systematic review and meta-analysis of experimental studies. Computers & Education, 227, 105224. https://doi.org/10.1016/j.compedu.2024.105224 [Google Scholar] [Crossref]

12. Fan, G., Liu, D., Zhang, R., & Pan, L. (2025). The impact of AI-assisted pair programming on student motivation, programming anxiety, collaborative learning, and programming performance: A comparative study with traditional pair programming and individual approaches. International Journal of STEM Education, 12(1), Article 16. https://doi.org/10.1186/s40594-025-00537-3 [Google Scholar] [Crossref]

13. Fawzy, A., Tahir, A., & Blincoe, K. (2026). From prompting to verification: How experience shapes vibe coding practices [Preprint]. arXiv. https://doi.org/10.48550/arXiv.2605.24521 [Google Scholar] [Crossref]

14. Fu, Y., Liang, P., Tahir, A., Li, Z., Shahin, M., Yu, J., & Chen, J. (2023). Security weaknesses of Copilot-generated code in GitHub projects: An empirical study [Preprint]. arXiv. https://doi.org/10.48550/arXiv.2310.02059 [Google Scholar] [Crossref]

15. Gama, K., Calegario, F., Jackson, V., Nolte, A., Morais, L. A., & Garcia, V. (2025). “Can you feel the vibes?”: An exploration of novice programmer engagement with vibe coding [Preprint]. arXiv. https://doi.org/10.48550/arXiv.2512.02750 [Google Scholar] [Crossref]

16. Gao, R., Tahir, A., Liang, P., Susnjak, T., & Khomh, F. (2025). A survey of bugs in AI-generated code [Preprint]. arXiv. https://doi.org/10.48550/arXiv.2512.05239 [Google Scholar] [Crossref]

17. Geruslu, V., Aliyeva, Z., & Tüzün, E. (2026). Factors influencing the quality of AI-generated code: A synthesis of empirical evidence [Preprint]. arXiv. https://doi.org/10.48550/arXiv.2603.25146 [Google Scholar] [Crossref]

18. GLM-5 Team. (2026). GLM-5: From vibe coding to agentic engineering [Preprint]. arXiv. https://doi.org/10.48550/arXiv.2602.15763 [Google Scholar] [Crossref]

19. Guo, R., Li, G., Miao, H., Pi, Z., & Xie, L. (2026). Impact of generative AI-assisted programming on the computational thinking of high school students. Frontiers in Psychology, 17, Article 1703177. https://doi.org/10.3389/fpsyg.2026.1703177 [Google Scholar] [Crossref]

20. Huang, R., Reyna, A., Lerner, S., Xia, H., & Hempel, B. (2025). Professional software developers don't vibe, they control: AI agent use for coding in 2025 [Preprint]. arXiv. https://doi.org/10.48550/arXiv.2512.14012 [Google Scholar] [Crossref]

21. Huesca, G., Martinez-Trevino, Y., Jiménez González, C. G., Cantú Delgado, D. A., Navarrete, C., Cedillo-Hernandez, A., & Quintero Meza, R. R. (2026). Teaching programming in the age of generative artificial intelligence: Learning gains and pedagogical integration in a higher education context. AI, 7(7), Article 248. https://doi.org/10.3390/ai7070248 [Google Scholar] [Crossref]

22. Hui, B., Yang, J., Cui, Z., Yang, J., Liu, D., Zhang, L., Liu, T., Zhang, J., Yu, B., Lu, K., Dang, K., Fan, Y., Zhang, Y., Yang, A., Men, R., Huang, F., Zheng, B., Miao, Y., Quan, S., ... Lin, J. (2024). Qwen2.5-Coder technical report [Preprint]. arXiv. https://doi.org/10.48550/arXiv.2409.12186 [Google Scholar] [Crossref]

23. Karpathy, A. [@karpathy]. (2025, February 2). There's a new kind of coding I call 'vibe coding', where you fully give in to the vibes, embrace exponentials, and forget that the code even exists [Post]. X. https://x.com/karpathy/status/1886192184808149383 [Google Scholar] [Crossref]

24. Kazemitabaar, M., Ye, R., Wang, X., Henley, A. Z., Denny, P., Craig, M., & Grossman, T. (2024). CodeAid: Evaluating a classroom deployment of an LLM-based programming assistant that balances student and educator needs. In Proceedings of the 2024 CHI Conference on Human Factors in Computing Systems. Association for Computing Machinery. https://doi.org/10.1145/3613904.3642773 [Google Scholar] [Crossref]

25. Kimi Team. (2026). Kimi K3: Open frontier intelligence [Preprint]. arXiv. https://doi.org/10.48550/arXiv.2607.24653 [Google Scholar] [Crossref]

26. Liao, J., Zhong, L., Zhe, L., Xu, H., Liu, M., & Xie, T. (2024). Scaffolding computational thinking with ChatGPT. IEEE Transactions on Learning Technologies, 17, 1628–1642. https://doi.org/10.1109/TLT.2024.3392896 [Google Scholar] [Crossref]

27. Licorish, S. A., Bajpai, A., Arora, C., Wang, F., & Tantithamthavorn, K. (2025). Comparing human and LLM generated code: The jury is still out! [Preprint]. arXiv. https://doi.org/10.48550/arXiv.2501.16857 [Google Scholar] [Crossref]

28. Manik, M. M. H. (2025). ChatGPT vs. DeepSeek: A comparative study on AI-based code generation [Preprint]. arXiv. https://doi.org/10.48550/arXiv.2502.18467 [Google Scholar] [Crossref]

29. Mishra, P., & Koehler, M. J. (2006). Technological pedagogical content knowledge: A framework for teacher knowledge. Teachers College Record, 108(6), 1017–1054. https://doi.org/10.1111/j.1467-9620.2006.00684.x [Google Scholar] [Crossref]

30. Page, M. J., McKenzie, J. E., Bossuyt, P. M., Boutron, I., Hoffmann, T. C., Mulrow, C. D., Shamseer, L., Tetzlaff, J. M., Akl, E. A., Brennan, S. E., Chou, R., Glanville, J., Grimshaw, J. M., Hróbjartsson, A., Lalu, M. M., Li, T., Loder, E. W., Mayo-Wilson, E., McDonald, S., ... Moher, D. (2021). The PRISMA 2020 statement: An updated guideline for reporting systematic reviews. BMJ, 372, Article n71. https://doi.org/10.1136/bmj.n71 [Google Scholar] [Crossref]

31. Pimenova, V., Fakhoury, S., Bird, C., Storey, M.-A., & Endres, M. (2025). Good vibrations? A qualitative study of co-creation, communication, flow, and trust in vibe coding [Preprint]. arXiv. https://doi.org/10.48550/arXiv.2509.12491 [Google Scholar] [Crossref]

32. Qwen Team. (2025). Qwen3 technical report [Preprint]. arXiv. https://doi.org/10.48550/arXiv.2505.09388 [Google Scholar] [Crossref]

33. Raihan, N., Siddiq, M. L., Santos, J. C. S., & Zampieri, M. (2025). Large language models in computer science education: A systematic literature review. In Proceedings of the 56th ACM Technical Symposium on Computer Science Education V. 1. Association for Computing Machinery. https://doi.org/10.1145/3641554.3701863 [Google Scholar] [Crossref]

34. Ren, S., Shen, X., Zhou, Y., Ng, D., & Raj, A. (2026). Chinese language is not more efficient than English in vibe coding: A preliminary study on token cost and problem-solving rate [Preprint]. arXiv. https://doi.org/10.48550/arXiv.2604.14210 [Google Scholar] [Crossref]

35. Rinja, D., Oliveira, E. A., López-Pernas, S., Saqr, M., Specht, M., & Misiejuk, K. (2026). Unpacking vibe coding: Help-seeking processes in student-AI interactions while programming [Preprint]. arXiv. https://doi.org/10.48550/arXiv.2604.27134 [Google Scholar] [Crossref]

36. Rogers, E. M. (2003). Diffusion of innovations (5th ed.). Free Press. [Google Scholar] [Crossref]

37. Smith, D. H., IV, & Zilles, C. (2024). Evaluating large language model code generation as an autograding mechanism for “Explain in Plain English” questions. In Proceedings of the 55th ACM Technical Symposium on Computer Science Education V. 2. Association for Computing Machinery. https://doi.org/10.1145/3626253.3635542 [Google Scholar] [Crossref]

38. Sobo, A., Mubarak, A., Baimagambetov, A., & Polatidis, N. (2024). Evaluating LLMs for code generation in HRI: A comparative study of ChatGPT, Gemini, and Claude. Applied Artificial Intelligence, 39(1), Article 2439610. https://doi.org/10.1080/08839514.2024.2439610 [Google Scholar] [Crossref]

39. Spracklen, J., Wijewickrama, R., Sakib, A. H. M. N., Maiti, A., Viswanath, B., & Jadliwala, M. (2025). We have a package for you! A comprehensive analysis of package hallucinations by code generating LLMs. In Proceedings of the 34th USENIX Security Symposium. USENIX Association. https://www.usenix.org/conference/usenixsecurity25/presentation/spracklen [Google Scholar] [Crossref]

40. Sun, D., Boudouaia, A., Zhu, C., & Li, Y. (2024). Would ChatGPT-facilitated programming mode impact college students' programming behaviors, performances, and perceptions? An empirical study. International Journal of Educational Technology in Higher Education, 21, Article 14. https://doi.org/10.1186/s41239-024-00446-5 [Google Scholar] [Crossref]

41. Thorgeirsson, S., Weidmann, T. B., & Su, Z. (2026). Computer science achievement and writing skills predict vibe coding proficiency [Preprint]. arXiv. https://doi.org/10.48550/arXiv.2603.14133 [Google Scholar] [Crossref]

42. Uddin, S. M. J. (2026). Is vibe coding the future? An empirical assessment of LLM generated codes for construction safety [Preprint]. arXiv. https://doi.org/10.48550/arXiv.2604.12311 [Google Scholar] [Crossref]

43. Umama, Danyaro, K. U., Nasser, M., Zakari, A., Abdullahi, S., Khanzada, A., Yakubu, M. M., & Shoaib, S. (2025). LLM-based code generation: A systematic literature review with technical and demographic insights. IEEE Access, 13, 194915–194939. https://doi.org/10.1109/ACCESS.2025.3631952 [Google Scholar] [Crossref]

44. Wermelinger, M. (2023). Using GitHub Copilot to solve simple programming problems. In Proceedings of the 54th ACM Technical Symposium on Computer Science Education V. 1 (pp. 172–178). Association for Computing Machinery. https://doi.org/10.1145/3545945.3569830 [Google Scholar] [Crossref]

45. Yetiştiren, B., Özsoy, I., Ayerdem, M., & Tüzün, E. (2023). Evaluating the code quality of AI-assisted code generation tools: An empirical study on GitHub Copilot, Amazon CodeWhisperer, and ChatGPT [Preprint]. arXiv. https://doi.org/10.48550/arXiv.2304.10778 [Google Scholar] [Crossref]

46. Zhao, S., Wang, D., Zhang, K., Luo, J., Li, Z., & Li, L. (2025). Is vibe coding safe? Benchmarking vulnerability of agent-generated code in real-world tasks [Preprint]. arXiv. https://doi.org/10.48550/arXiv.2512.03262 [Google Scholar] [Crossref]

47. Zhu, Q., Cao, J., Lu, Y., Lin, H., Han, X., Sun, L., & Cheung, S.-C. (2024). DOMAINEVAL: An auto-constructed benchmark for multi-domain code generation [Preprint]. arXiv. https://doi.org/10.48550/arXiv.2408.13204 [Google Scholar] [Crossref]

48. Ziegler, A., Kalliamvakou, E., Li, X. A., Rice, A., Rifkin, D., Simister, S., Sittampalam, G., & Aftandilian, E. (2024). Measuring GitHub Copilot's impact on productivity. Communications of the ACM, 67(3), 54–63. https://doi.org/10.1145/3633453 [Google Scholar] [Crossref]

Frontier AI Model Output for Educational Application Development via Vibe-Coding: A Systematic Literature Review Comparing Western and Eastern AI Models

Authors

Article Information

Publication Timeline

Abstract

Keywords

Downloads

References

Metrics

Views & Downloads

Similar Articles