COMPARISON OF TELEGRAM BOT FOR TEXT GENERATION BUILT ON OPENAI AND GEMINIAI

Abstract

The study examines the problem of developing and conducting a comparative analysis of modern generative language models, using OpenAI GPT and Gemini 2.5 Pro integrated into a Telegram bot for automated text generation. The relevance of the research is driven by the increasing demand for fast, flexible, and high-quality content creation tools capable of operating in real time. Within the work, the requirements for the software tool were defined, the architecture of the Telegram bot was developed, the principles of its operation were determined, and the interaction with the APIs of both models was implemented. Special attention is given to testing the quality of generated texts, response speed, model stability, and their behavior under load. Experimental results indicate that Gemini 2.5 Pro demonstrates significantly higher generation speed and more precise interpretation of user queries, whereas OpenAI GPT excels in creative tasks, stylistic naturalness, and flexibility of formulations. Both models show strong performance across multiple text genres, though they differ in structural coherence, logic, and context retention.The results of the study show that neither model has an absolute advantage; the choice depends on specific tasks and requirements for style, accuracy, or speed of response. The developed Telegram bot confirms the effectiveness of integrating generative models into practical systems and demonstrates adaptability to various usage scenarios. The findings can be applied to further optimization of generative systems, model selection for specific applications, and the development of intelligent AI-based services.

Ähnliche Arbeiten