Boogu-Image-0.1 Tops HuggingFace Daily With 727 Upvotes: Frontier-Adjacent Image Generation Trained for ~$400K

arXiv / HuggingFace Daily Papers·high signal

A 33-author team released a unified multimodal understanding-and-generation model under Apache 2.0 — weights, code, and recipes — trained on only 208.62 million unique images at a theoretical training cost of roughly $400K. It ships in four variants (Base, Turbo, Edit, Edit-Turbo) covering text-to-image, fast inference, instruction-based editing, and bilingual Chinese-English text rendering, and claims it 'consistently matches or surpasses other open-source models across standard benchmarks' while approaching closed-source systems. The 727-upvote margin over the next paper (154) makes this the runaway paper of the day; the cost figure is the real story for builders.