RefSeq sequences are for organisms that may or may not have completed genome. Refseq under the genomes section is for complete genomes.

Refseq:

The NCBI RefSeq project is an ongoing effort to provide a curated, non-redundant collection of reference sequences, representative of the central dogma, for each major organism. The full release incorporates genomic, transcript, and protein data available at the time of each release.

Genomes:

Sequence data is provided for all single organism genome assemblies that are included in NCBI's Assembly resource www.ncbi.nlm.nih.gov/assembly/)

refseq: content includes assembled genome sequence and RefSeq annotation data. All prokaryotic and eukaryotic RefSeq genomes have annotation. RefSeq annotation data may be calculated by NCBI annotation
pipelines or propagated from the GenBank submission. The RefSeq directory area includes fewer organisms than the GenBank directory area because not all genome assemblies are selected for the RefSeq project.