SciDraw-6K: A Multilingual Scientific Illustration Dataset Generated by Google Gemini

SciDraw-6K is a curated dataset of 6,291 scientific illustrations synthesized by Google Gemini image-generation models, each paired with prompts in 11 languages (English, Simplified and Traditional Chinese, Japanese, Korean, German, French, Spanish, Brazilian Portuguese, Italian, and Russian). Images span eight broad scientific categories: biomedical, chemistry, materials, electronics, environment, AI systems, physics, and a residual "other" bucket covering long-tail disciplines. Images are produced primarily by the gemini-2.5-flash-image and gemini-3-pro-image-preview model families. Unlike general-purpose text-to-image corpora, SciDraw-6K is purpose-built for the scientific-illustration genre: schematic diagrams, mechanism figures, table-of-contents graphics, and conceptual posters. The dataset supports multilingual T2I research, domain-adapted diffusion fine-tuning, and prompt-engineering studies for scientific visualization. Full dataset (including all image files, ~19 GB) is hosted on Hugging Face:https://huggingface.co/datasets/SciDrawAI/SciDraw-6K This Zenodo record provides a citable DOI and archives the dataset metadata (metadata.parquet and metadata.jsonl), along with a README documenting the schema, construction pipeline, and intended uses. The dataset powers sci-draw.com, a public scientific drawing service: https://sci-draw.com