U-Mamba is a novel network architecture designed for biomedical image segmentation, addressing the limitations of both Convolutional Neural Networks (CNNs) and Transformers in handling long-range dependencies. Inspired by State Space Sequence Models (SSMs), U-Mamba integrates the local feature extraction capabilities of CNNs with the long-range dependency modeling strengths of SSMs. The architecture includes a self-configuring mechanism that allows it to adapt to various datasets without manual intervention. Extensive experiments on four diverse tasks—3D abdominal organ segmentation in CT and MRI images, instrument segmentation in endoscopy images, and cell segmentation in microscopy images—show that U-Mamba outperforms state-of-the-art CNN- and Transformer-based networks. U-Mamba's linear scaling in feature size and efficient long-range dependency modeling make it a promising candidate for future advancements in biomedical image analysis. The code, models, and data are publicly available at <https://wanglab.ai/u-mamba.html>.U-Mamba is a novel network architecture designed for biomedical image segmentation, addressing the limitations of both Convolutional Neural Networks (CNNs) and Transformers in handling long-range dependencies. Inspired by State Space Sequence Models (SSMs), U-Mamba integrates the local feature extraction capabilities of CNNs with the long-range dependency modeling strengths of SSMs. The architecture includes a self-configuring mechanism that allows it to adapt to various datasets without manual intervention. Extensive experiments on four diverse tasks—3D abdominal organ segmentation in CT and MRI images, instrument segmentation in endoscopy images, and cell segmentation in microscopy images—show that U-Mamba outperforms state-of-the-art CNN- and Transformer-based networks. U-Mamba's linear scaling in feature size and efficient long-range dependency modeling make it a promising candidate for future advancements in biomedical image analysis. The code, models, and data are publicly available at <https://wanglab.ai/u-mamba.html>.