OMG-Seg is a unified segmentation model designed to handle a wide range of visual segmentation tasks, including image semantic, instance, and panoptic segmentation, as well as their video counterparts, open-vocabulary settings, prompt-driven, interactive segmentation, and video object segmentation. This is the first model to unify these tasks in one framework, achieving competitive performance across various datasets. OMG-Seg employs a transformer-based encoder-decoder architecture with task-specific queries and outputs, enabling efficient and effective handling of multiple tasks while reducing computational and parameter overhead. The model uses a shared encoder-decoder architecture, where a single query can represent different types of masks, IDs, and visual prompts, allowing for streamlined processing with a shared decoder. Through co-training on combined image and video datasets, OMG-Seg demonstrates superior performance compared to specialized models and other unified models, making it a versatile and efficient solution for diverse segmentation tasks.OMG-Seg is a unified segmentation model designed to handle a wide range of visual segmentation tasks, including image semantic, instance, and panoptic segmentation, as well as their video counterparts, open-vocabulary settings, prompt-driven, interactive segmentation, and video object segmentation. This is the first model to unify these tasks in one framework, achieving competitive performance across various datasets. OMG-Seg employs a transformer-based encoder-decoder architecture with task-specific queries and outputs, enabling efficient and effective handling of multiple tasks while reducing computational and parameter overhead. The model uses a shared encoder-decoder architecture, where a single query can represent different types of masks, IDs, and visual prompts, allowing for streamlined processing with a shared decoder. Through co-training on combined image and video datasets, OMG-Seg demonstrates superior performance compared to specialized models and other unified models, making it a versatile and efficient solution for diverse segmentation tasks.