AbstractGINS is a key component of eukaryotic replicative forks and is composed of four subunits (Sld5, Psf1, Psf2, Psf3). To explain the discrepancy between structural data from crystallography and electron microscopy (EM), we show that GINS is a compact tetramer in solution as observed in crystal structures, but also forms a double-tetrameric population, detectable by EM. This may represent an intermediate step towards the assembly of two replicative helicase complexes at origins, moving in opposite directions within the replication bubble. Reconstruction of the double-tetrameric form, combined with small-angle X-ray scattering data, allows the localisation of the B domain of the Psf1 subunit in the free GINS complex, which was not visible in previous studies and is essential for the formation of a functional replication fork.