This study conducts a comprehensive comparison of Whisper Automatic Speech Recognition (ASR) models–medium, turbo, and large–across 15 Indic languages. We have taken the dataset from IndicTTS. For each language, we evaluated model performance, measuring Word Error Rate (WER), Character Error Rate (CER), and Phoneme Error Rate (PER). Significantly, all whisper variants struggled to recognise specific low-resource languages, including Dogri, Maithili, Manipuri, and Rajasthani, underscoring existing gaps in multilingual coverage for the ASR task. Consistent with expectations, the larger whisper models demonstrated superior accuracy over their smaller models. Most languages, such as Hindi and Tamil, are the best-performing languages among all 15 Indic languages. Our findings offer actionable insights into the strengths and current limitations of whisper models, reinforcing the need for further research in supporting underrepresented languages.