Ray-finned fishes – especially teleosts like zebrafish – are used to investigate the genomic basis of vertebrate development and evolution. However, teleosts are derived from a whole genome duplication (WGD) that had major impact on their genome and gene functions. Together with the two earlier vertebrate WGDs, this complicates macroevolutionary comparisons among vertebrates: WGDs led to lineage-specific genome reshuffling and gene losses, obscuring the distinction of orthologs vs. paralogs and hiding the origins of gene functions and developmental processes. We show that so-called ‘ancient’ holostean fishes (gars and bowfin), teleosts’ sister lineage, have very informative genomes and body plans providing unique opportunities for comparative investigations. Holosteans serve both as ‘unduplicated’ outgroup to the 30,000 teleost species as well as an outgroup to lobe-finned vertebrates including 30,000 tetrapod species. Using examples from diverse developmental pathways and processes, we show that comparative genomic, developmental, transcriptomic, and epigenomic analyses with holosteans as “bridge species” are indispensable for connecting the often disparate sets of genes, gene regulatory elements, and morphologies among distant vertebrate lineages. The genome of our main holostean model species, the spotted gar (Lepisosteus oculatus), is representative of the bony vertebrate ancestor and retained numerous genes differentially lost in other lineages. The ‘evolutionary inertia’ of holosteans facilitates to find deep homology of regulatory elements across vertebrate lineages. Rearing gars in the laboratory, we developmentally test hypotheses about the evolutionary origins of vertebrate gene functions. Holosteans are thus integrative Evo-Devo models that illuminate vertebrate biology.