Interleaved multimodal generation enables capabilities beyond unimodal generation models, such as step-by-step instructional guides, visual planning, and generating visual drafts for reasoning. However, the quality of existing interleaved generation models under general instructions remains limited by insufficient training data and base model capacity. We present DuoGen, a general-purpose interleaved generation framework that systematically addresses data curation, architecture design, and evaluation. On the data side, we build a large-scale, high-quality instruction-tuning dataset by combining multimodal conversations rewritten from curated raw websites, and diverse synthetic examples covering everyday scenarios. Architecturally, DuoGen leverages the strong visual understanding of a pretrained multimodal LLM and the visual generation capabilities of a diffusion transformer (DiT) pretrained on video generation, avoiding costly unimodal pretraining and enabling flexible base model selection. A two-stage decoupled strategy first instruction-tunes the MLLM, then aligns DiT with it using curated interleaved image-text sequences. Across public and newly proposed benchmarks, DuoGen outperforms prior open-source models in text quality, image fidelity, and image-context alignment, and also achieves state-of-the-art performance on text-to-image and image editing among unified generation models. Data and code will be released at https://research.nvidia.com/labs/dir/duogen/.
@article{arxiv.2602.00508,
title = {DuoGen: Towards General Purpose Interleaved Multimodal Generation},
author = {Min Shi and Xiaohui Zeng and Jiannan Huang and Yin Cui and Francesco Ferroni and Jialuo Li and Shubham Pachori and Zhaoshuo Li and Yogesh Balaji and Haoxiang Wang and Tsung-Yi Lin and Xiao Fu and Yue Zhao and Chieh-Yun Chen and Ming-Yu Liu and Humphrey Shi},
journal= {arXiv preprint arXiv:2602.00508},
year = {2026}
}