DNIPRO is a longitudinal corpus of 246,229 news articles documenting the Russo-Ukrainian war from February 2022 to August 2024. The dataset captures competing geopolitical narratives from 11 media outlets across five nation-states (Russia, Ukraine, U.S., U.K., China) in three languages (English, Russian, Mandarin Chinese). The corpus includes comprehensive metadata, named entity annotations, sentiment scores for key actors and events, and topical framing labels based on a nine-category schema. All annotations are validated by human annotators. The dataset is designed to support research in computational journalism, media framing analysis, information warfare studies, and cross-lingual narrative analysis. Note: Due to potential copyright restrictions, this release contains metadata and annotations only. Full article text must be retrieved via the provided URLs. All data collection complied with publishers' terms of service. For detailed methodology, see the accompanying preprint (arXiv:2601.16309), to be published at the Fifteenth biennial Language Resources and Evaluation Conference (LREC) 2026. Keywords: news corpus, media framing, Russo-Ukrainian war, multilingual, geopolitical narratives, information warfare, sentiment analysis, stance detection