Spectrum sensing is a pivotal function in cognitive radio systems, enabling dynamic spectrum access by identifying underutilized frequency bands. Traditional spectrum sensing methodologies, such as energy detection and matched filtering, often struggle with low signal-to-noise ratio (SNR) conditions and require prior knowledge of the primary user signals. Recent advances in deep learning have shown promise in addressing these limitations by learning complex patterns from data without requiring explicit signal models. In this article, we propose a novel spectrum sensing methodology based on Transformer architecture, a deep learning model renowned for its ability to capture long-range dependencies and process sequential data efficiently. The proposed approach leverages the self-attention mechanism of Transformers to enhance detection accuracy and robustness under varying channel conditions. Experimental results demonstrate that the Transformer-based model outperforms traditional and existing deep learning methods, particularly in challenging and realistic environments such as Rayleigh channel. This study highlights the potential of Transformers for sensing the spectrum efficiently in cognitive radio networks, paving the way for more intelligent and adaptive wireless communication systems.