<![CDATA[The massive adoption of Large Language Models (LLMs) across sectors demands an in-depthevaluation of the systems' intrinsic capabilities, particularly in responding to ambiguousinstructions. This study presents a structured comparative analysis of three leading LLMarchitectures ChatGPT, DeepSeek, and Gemini focusing on factual reliability and computationalefficiency (inference latency) as a direct consequence of prompt-ambiguity manipulation. Themethodology adopts a Black-Box Testing approach combined with quantitative analysis,systematically applied to 100 prompts classified into eight complexity categories. Factual reliabilitywas measured using a TF-IDF-based Cosine Similarity metric against ground truth, while latencywas recorded through end-to-end stopwatch measurement and processed with Python scripts onGoogle Colab. Results show that Gemini achieved the highest average accuracy (66.91%), followedby DeepSeek (64.89%) and ChatGPT (62.10%), while DeepSeek led in average inference speed(2.84 seconds) owing to its Mixture-of-Experts architecture. Pearson correlation tests for all threemodels produced negative coefficients (ChatGPT −0.200; Gemini −0.194; DeepSeek −0.030),disproving the traditional speed–accuracy trade-off hypothesis. These findings are expected to serveas an empirical compass for users selecting models with precision and as evaluative literature fordevelopers building more robust and efficient AI systems.]]>
Copyrights © 2026
