Flaviviruses are arthropod-borne viruses found worldwide and are responsible for significant human and veterinary diseases, including dengue, Zika, and West Nile fever. Some flaviviruses are insect specific and replicate only in mosquitoes. We report a genetically divergent group of insect-specific flaviviruses from mosquitoes that do not replicate in arthropod cell lines or heterologous species, exhibiting unprecedented specialization for their host species. Determination of the complete sequences of the RNA genomes of three of these viruses, Karumba virus (KRBV), Haslams Creek virus, and Mac Peak virus (McPV), that are found in high prevalence in some mosquito populations and detection of virus-specific proteins, replicative double-stranded RNA, and small interfering RNA responses in the host mosquito species provided strong evidence of a functional replicating virus in the mosquito midgut. Analysis of nucleotide composition in the KRBV and McPV sequences also revealed a pattern consistent with the virus evolving to replicate only in insects. These findings represent a significant advance in our knowledge of mosquito-borne flavivirus ecology, host restriction, and evolution. Flaviviruses like dengue, Zika, or West Nile virus infect millions of people each year and are transmitted to humans via infected-mosquito bites. A subset of flaviviruses can only replicate in the mosquito host, and recent studies have shown that some can interfere with pathogenic flaviviruses in mosquitoes and limit the replication and transmission of the latter. The insect-specific flaviviruses (ISFs) reported here form a new mosquito-associated clade separate from the - and -associated ISF clades. The identification of distinct clades for each mosquito genus provides new insights into the evolution and ecology of flaviviruses. One of these viruses was shown to replicate in the midgut of the mosquito host and exhibit the most specialized host restriction reported to date for ISFs. Understanding this unprecedented host restriction in ISFs could help identify the mechanisms involved in the evolution of flaviviruses and their emergence as mosquito-borne pathogens.