Context. High-precision data from AMS-02 on Li, Be, and B provide the best constraints on Galactic cosmic-ray transport parameters. Aims. We re-evaluate the impact of Fe fragmentation on the Li, Be, and B modelling. We discuss the consequences on the transport parameter determination and reassess whether a primary source of Li is needed to match AMS-02 data. Methods. We renormalised several cross-section parametrisations to existing data for the most important reactions producing Li, Be, and B. We used the usine code with these new cross-section sets to re-analyse Li/C, Be/C, and B/C AMS-02 data. Results. We built three equally plausible cross-section sets. Compared to the initial cross-section sets, they lead to an average enhanced production of Li (∼ 20 − 50%) and Be (∼ 5 − 15%), while leaving the B flux mostly unchanged. In particular, Fe fragmentation is found to contribute to up to 10% of the Li and Be fluxes. Used in the combined analysis of AMS-02 Li/C, Be/C, and B/C data, the fit is significantly improved, with an enhanced diffusion coefficient (∼ 20%). The three updated cross-section sets are found to either slightly undershoot or overshoot the Li/C and B/C ratios: this strongly disfavours evidence for a primary source of Li in cosmic rays. We stress that isotopic cosmic-ray ratios of Li (and to a lesser extent Be), soon to be released by AMS-02, are also impacted by the use of these updated sets. Conclusions. Almost no nuclear data exist for the production of Li and B isotopes from Ne, Mg, Si, and Fe, whereas these reactions are estimated to account for ∼ 20% of the total production. New nuclear measurements would be appreciated and help to better exploit the high-precision AMS-02 cosmic-ray data.