Reading long documents, newspapers and textbooks can be generally a time-consuming process and is complex in nature. Summarization techniques play a vital role for understanding the text in simplest forms of optimal fashion. In research areas creating summaries for Research / Scientific Articles in order to get a clear picture of what the article is about, what information it contains, and so on. To accomplish this, we present SIMCLS, a conceptual framework for summarization in abstractive manner that can combine the gap between the learning objective and evaluation metrics caused by the presently dominant sequence-to- sequence learning technique by articulating text generation as a reference-free evaluation problem approach-driven by contrastive learning. Experiment analysis interpret that SimCLS can significantly enhance the evaluation of existing top-performing models with minor modifications over existing top-scoring systems. Abstractive Summarization is a type of summarization that uses rephrasing or creating new words to generate novel sentences and thus summaries for research articles, as opposed to Extractive Summarization, which uses extraction to extract important words and then generate a summary. Contrastive learning, a self- supervised, task independent deep learning technique, allows a model to learn about data even in the absence of labels. SimCLS is a simple framework that employs abstractive summarization and contrastive learning. This is used to generate summaries of research articles.