The paper introduces ASTRAL, a fast and statistically consistent method for estimating species trees from multiple genes. ASTRAL is designed to address the challenges posed by incomplete lineage sorting (ILS) and other factors that can cause gene trees to differ from species trees. The method is based on the multi-species coalescent model and can handle large datasets with thousands of genes. ASTRAL is compared to other leading coalescent-based methods, such as MP-EST and BUCKy, as well as summary methods like MRP and the greedy consensus. The results show that ASTRAL is more accurate than these methods, especially under moderate to high levels of ILS. ASTRAL is also faster than other methods, making it suitable for genome-scale analyses. The paper includes experimental results on both simulated and biological datasets, demonstrating ASTRAL's effectiveness in resolving complex phylogenetic relationships.The paper introduces ASTRAL, a fast and statistically consistent method for estimating species trees from multiple genes. ASTRAL is designed to address the challenges posed by incomplete lineage sorting (ILS) and other factors that can cause gene trees to differ from species trees. The method is based on the multi-species coalescent model and can handle large datasets with thousands of genes. ASTRAL is compared to other leading coalescent-based methods, such as MP-EST and BUCKy, as well as summary methods like MRP and the greedy consensus. The results show that ASTRAL is more accurate than these methods, especially under moderate to high levels of ILS. ASTRAL is also faster than other methods, making it suitable for genome-scale analyses. The paper includes experimental results on both simulated and biological datasets, demonstrating ASTRAL's effectiveness in resolving complex phylogenetic relationships.