Direct evidence of quantum coherence in a single-molecule magnet in frozen solution is reported with coherence times as long as T2 = 630 ± 30 ns. We can strongly increase the coherence time by modifying the matrix in which the single-molecule magnets are embedded. The electron spins are coupled to the proton nuclear spins of both the molecule itself and interestingly, also to those of the solvent. The clear observation of Rabi oscillations indicates that we can manipulate the spin coherently, an essential prerequisite for performing quantum computations.PACS numbers: 75.30.Gw, 75.50.Xx, The key concept in quantum information processing is that a quantum bit (qubit) may be not just 0 or 1, as in ordinary computer bits, but an arbitrary superposition of 0 and 1. This means that any two-level system, that can be put into a superposition state, is a qubit candidate [1]. The required superposition state is created by electromagnetic radiation pulses with a frequency corresponding to the energy splitting between the two levels ( Fig. 1c). The contribution of each of the two levels to the superposition state has a cyclic dependence on the pulse length, leading to so-called Rabi oscillations [1]. The observation of such oscillations is a proof-of-principle for the viability of performing quantum computations with a particular system. Quantum computers will probably not be realized from single atoms but will most likely utilize solid state devices, such as superconducting junctions, semiconductor structures, or molecular magnets [1,2]. For these large systems the quantum coherence decays fast, which drastically shortens the time available for quantum computation. Molecular magnets have been considered as qubits because they can be easily organized into large-scale ordered arrays by surface self-assembly [? ], and because they possess excited electronic-spin states required for two-qubit gate operations [5,6]. Single-molecule magnets (SMMs) are exchange-coupled clusters with highspin ground states [3]. The Ising-type anisotropy creates an energy barrier toward magnetization relaxation [ Fig. 1(b)], and many fascinating quantum phenomena have been observed in these systems, such as quantum tunnelling of the magnetization and quantum phase interference [3]. The large splitting of the two lowest states of SMMs in zero field (in principle) allows performing coherent spin-manipulations without external magnetic field, which simplifies any practical implementation. SMMs have also been proposed for the implementation of Grover's algorithm [2] allowing numbers between 0 and 2 2S−2 to be stored in a single molecule.The long coherence time is a crucial first step towards successful implementation of SMMs as qubits [1]. Therefore, recent years have seen a great deal of activity in trying to determine the quantum coherence times in SMMs, which was estimated to be of the order of 10 ns [7,8,9,10]. In several cases, energy gaps between superposition states have been reported that are larger than the expected decoherence energy scale [...