We introduce a soft-detection variant of Guessing Random Additive Noise Decoding (GRAND) called Quantized GRAND (QGRAND) that can efficiently decode any moderate redundancy block-code in an algorithm that is suitable for highly parallelized implementation in hardware. QGRAND can avail of any level of quantized soft information, and is shown to provide near maximum likelihood decoding performance when provided with five or more bits of soft information per received bit.