I know of several MBA programmes that teach systems thinking as action learning projects. Some benchmarking between them should be possible.
In my experience, systemic action learning involving projects works very well, while action learning directed at functional improvements can be problematical.
To explain: in terms of the Biomatrix Systems Approach, both a project and function are activity systems and organised according to the same framework and principles. The difference between them is that projects are a temporary part of a system and come to an end when their aim is achieved, while a function is an inherent and ongoing part of an entity system (e.g. the nutrition or thinking function of a person; the marketing, financial or production function of an organisation; the education or energy provision function of a society). As the entity changes, so do its functions; and vice versa.
In a functional context, problems are typically co-produced by the function itself, by other functions within the organisation and by the organisation as a whole (besides the functional environment). Thus an isolated action learning project aimed at improving a function can achieve only the elimination of the problems that arise from within the system. To eliminate the other problems and achieve a systemic functional transformation would require changes in other functions and the organisation as a whole. This is only possible in the context of an organisation transformation programme such as the Biomatrix Organisation Development Programme. It cannot be achieved in an isolated functional intervention.
By comparison, the scope of projects is more limited. They typically address a specific problem or need of a system by providing a product or service to it. Examples are a building project, marketing proposal or new product development.
Benchmarking in terms of specific outcomes of the projects (type II properties) is not really possible, except maybe in the broadest terms (e.g. implementation within budget and time). Benchmarking of type I properties such as adhering to the prescribed steps of a procedure could be useful.
Action learning programmes can also be set up to experience the difference between conventional and systemic project design and implementation. For example, for several years I presented a module on systems thinking in the context of an action learning programme which involved projects prescribed by the organisation. It was the last module in the programme and most teams were in the process of finalising their project design and implementation plan. After exposure to systems thinking they realised how much they had not considered in their conventional project design and that they need to iterate through it again and make some significant changes and additions. Every year we got the advice from students to put systems thinking as the first module. We explained that they still lacked detailed knowledge at the beginning, that integration is only possible at the end and that they would have missed fundamental learning (i.e. that systemic project management is more effective than conventional project management). In all the years, all projects were implemented, with the exception of one, whose team refused to rework their project design after the systems thinking module.