Inadequate knowledge regarding specific local and interactive behavior of slender high‐strength steel (HSS) hollow sections presents an obstacle in implementing these sections in the construction practice. The current approach in the standard is simplified and offers overly conservative results. Dealing with non‐standard cross‐sections only expands on these difficulties. This paper contributes to the on‐going RFCS project "HOLLOSSTAB", which has been dealing with the aforementioned issues. The project design proposal is based on the "Overall interaction concept" (OIC), and a new set of design rules for hollow sections is currently being developed. In this concept, linear buckling analysis (LBA) is used to obtain the slenderness of the member, and Geometrically and Materially Non‐linear Imperfection Analysis (GMNIA) is used to determine an "overall" buckling reduction factor. Extensive experimental tests are used to validate the method, by correlating the experimental results with numerical test results (GMNIA‐real) and statistically analyzing them. The focus of the tests in this paper is the use of high‐strength steel, ranging from S500 to S890, used for rectangular hollow‐sections and hexagonal‐hollow sections. The state‐of‐the‐art measuring tools available, offer the possibility of precise reverse‐engineering process, creating numerical models based on experimental test, which allows an accurate prediction of the ultimate load capacity of the specimens. Digital image correlation (DIC) allows a review of the buckling shape, comparing numerical model with real shape of tested specimen. The paper aims to validate the numerical models, as well as validity of assumptions on imperfection amplitude. [ABSTRACT FROM AUTHOR]