This work compares two different X-ray flat-panel detectors for its use in high-speed, cone-beam CT applied to small-animal imaging. The main differences between these two devices are the scintillators and the achievable frame rate. Both devices have been tested in terms of system linearity, sensitivity, resolution, stability and noise properties, taking into account the different timing schemes for each one of them and the mandatory corrections on the raw data. Tomographic scans have been carried out using both detectors to evaluate its final performance as well as the delivered dose needed to achieve similar quality scans. An experimental cone-beam CT test-bench has been designed and implemented to perform the different measurements. It uses a micro-focus X-ray source and a rotating stage where the samples are placed. A modified FDK algorithm has been used to reconstruct the acquired data. Both detectors show similar results for pixel linearity and stability measurements, and their noise levels are comparable. The resolution and sensitivity features are better for the direct grown scintillator detector (9 Ipmm vs. 6 Ipmm, and ∼4 times more sensitive for similar delivered dose). Since tomographic reconstructed images for the higher frame-rate detector show acceptable quality, it can be used to implement a faster system for high-speed acquisition techniques like, for example, dynamic imaging or gated protocols.