The study of energy conversion at microscopic scales is central in nonequilibrium thermodynamics, such as physics, chemistry, biology, and quantum technologies. In this talk, I’ll discuss topics in stochastic thermodynamics of collective systems. Unlike typical phase transitions characterized by a single critical point, we show that the interplay between two temperatures and biased forces can split the order-disorder transition into different two points, depending on the dominant ordered state. We also show that, under general conditions, the heat engine is delimited by the minimal power fluctuations and entropy production, which together delimit its optimal performance—achieved when these conditions are fully satisfied. Finally, we explore a multi-objective optimization framework using Pareto fronts, revealing that while few-variable optimization leads to convex fronts, increasing complexity and non conservative forces result in concave fronts