Images in Sentences: Scaling Interleaved Instructions for Unified Visual Generation

TL;DR

INSET is a unified model embedding images as native tokens within text, improving multi-image instruction understanding and generation by leveraging transformer locality and a large synthetic dataset.

Contribution

The paper introduces INSET, a novel model that seamlessly integrates images into textual instructions and a scalable data engine for training on 15 million synthetic interleaved samples.

Findings

INSET outperforms state-of-the-art in multi-image consistency and text alignment.

Performance gaps increase with input complexity.

The approach extends naturally to multimodal image editing.

Abstract

While recent advancements in multimodal language models have enabled image generation from expressive multi-image instructions, existing methods struggle to maintain performance under complex interleaved instructions. This limitation stems from the structural separation of images and text in current paradigms, which forces models to bridge difficult long-range dependencies to match descriptions with visual targets. To address these challenges, we propose \texttt{I}mages i\texttt{N} \texttt{SE}n\texttt{T}ences (\textit{a.k.a}, INSET), a unified generation model that seamlessly embeds images as native vocabulary within textual instructions. By positioning visual features directly at their corresponding semantic slots, INSET leverages the contextual locality of transformers for precise object binding, effectively treating images as dense, expressive language tokens. Furthermore, we introduce a scalable data engine that synthesizes 15M high-quality interleaved samples from standard image and video datasets, utilizing VLMs and LLMs to construct rich, long-horizon sequences. Evaluation results on InterleaveBench demonstrate that INSET significantly outperforms state-of-the-art methods in multi-image consistency and text alignment, with performance gaps widening as input complexity increases. Beyond standard generation, our approach inherently extends to multimodal image editing, integrating visual content as part of the instruction to facilitate highly expressive and creative visual manipulations.