This work introduces a method that focuses on enhancing resource allocation in cloud computing environments by considering Quality of Service (QoS) factors. Since resource allocation plays a crucial role in determining the QoS of cloud services, it is important to consider indicators like response time, throughput, waiting time, and makespan. The primary difficulty in cloud computing lies in resource allocation, which can be tackled by proposing a novel algorithm known as Modified Fire Hawks Gazelle Optimization (MFHGO). The proposed approach involves the hybridization of the modified fire hawks algorithm with gazelle optimization to facilitate efficient resource allocation. It aims to optimize several objectives, such as resource utilization, degree of imbalance, completion time, throughput, relative error, and response time. To achieve this, an optimal resource allocation is achieved using the Partitioning around K-medoids (PAKM) clustering approach. The proposed model extends the K-means clustering method. For simulation purposes, the GWA-T-12 Bitbrains dataset is utilized, while the JAVA tool is employed for exploratory analysis. The effectiveness of the proposed resource allocation and clustering approach is demonstrated by comparing it with existing schemes. The proposed work's makespan is 1.45 seconds for 50 tasks, 3.6 seconds for 100 tasks, 3.67 seconds for 150 tasks, and 5.34 seconds for 200 jobs. As a result, the proposed model achieves the smallest makespan value when compared to the previous approaches. The proposed work yielded response times of 105ms for a task length of 100, 376ms for 200, 555ms for 300, 624ms for 400, and 1014ms for 500. These results indicate that the proposed model outperforms current approaches by achieving a faster response time and also attains a bandwidth utilization of 0.80%, 0.90%, and 0.97% for 4, 6, and 16 tasks, respectively, indicating better bandwidth utilization than the other approaches.