The paper introduces nnMamba, a novel architecture designed for 3D biomedical image analysis, focusing on segmentation, classification, and landmark detection. nnMamba integrates the strengths of Convolutional Neural Networks (CNNs) and State Space Models (SSMs) to effectively capture long-range dependencies in high-dimensional medical images. The key contributions include the Mamba-In-Convolution with Channel-Spatial Siamese learning (MICCSS) block, which models long-range relationships between voxels, and channel-scaling and channel-sequential learning methods for dense prediction and classification tasks. Extensive experiments on six datasets demonstrate nnMamba's superior performance compared to state-of-the-art methods, making it a robust solution for 3D medical image analysis. The code and data are available for public use.The paper introduces nnMamba, a novel architecture designed for 3D biomedical image analysis, focusing on segmentation, classification, and landmark detection. nnMamba integrates the strengths of Convolutional Neural Networks (CNNs) and State Space Models (SSMs) to effectively capture long-range dependencies in high-dimensional medical images. The key contributions include the Mamba-In-Convolution with Channel-Spatial Siamese learning (MICCSS) block, which models long-range relationships between voxels, and channel-scaling and channel-sequential learning methods for dense prediction and classification tasks. Extensive experiments on six datasets demonstrate nnMamba's superior performance compared to state-of-the-art methods, making it a robust solution for 3D medical image analysis. The code and data are available for public use.