The paper introduces MambaMIR, a Mamba-based model for medical image reconstruction and uncertainty estimation, along with its GAN variant, MambaMIR-GAN. MambaMIR leverages the advantages of the original Mamba model, such as linear complexity, global receptive fields, and dynamic weights, to effectively handle long sequences and preserve detailed information in medical images. The proposed model includes an innovative arbitrary-mask mechanism, which provides randomness for Monte Carlo-based uncertainty estimation by randomly masking redundant scan sequences. Experiments on various medical image reconstruction tasks, including fast MRI and sparse-view CT (SVCT), demonstrate that MambaMIR and MambaMIR-GAN achieve comparable or superior reconstruction results compared to state-of-the-art methods. Additionally, the estimated uncertainty maps offer insights into the reliability of the reconstructed images. The code for MambaMIR is publicly available at https://github.com/ayanglab/MambaMIR.The paper introduces MambaMIR, a Mamba-based model for medical image reconstruction and uncertainty estimation, along with its GAN variant, MambaMIR-GAN. MambaMIR leverages the advantages of the original Mamba model, such as linear complexity, global receptive fields, and dynamic weights, to effectively handle long sequences and preserve detailed information in medical images. The proposed model includes an innovative arbitrary-mask mechanism, which provides randomness for Monte Carlo-based uncertainty estimation by randomly masking redundant scan sequences. Experiments on various medical image reconstruction tasks, including fast MRI and sparse-view CT (SVCT), demonstrate that MambaMIR and MambaMIR-GAN achieve comparable or superior reconstruction results compared to state-of-the-art methods. Additionally, the estimated uncertainty maps offer insights into the reliability of the reconstructed images. The code for MambaMIR is publicly available at https://github.com/ayanglab/MambaMIR.