Common forms for emerging teams in the open-source AI community to release results include: publishing repositories on GitHub (including model weights, training scripts, inference code, and README documentation), releasing technical reports or model cards, building interactive demos or API interfaces, and submitting preprint papers on arXiv.