Despite the importance of silicon (Si) as benefecial nutrient for many plants, including economically-important cereals, the reactivity of various Si pools in soils (silicate minerals, amorphous compounds, phytoliths, organic litter) is not fully quantified which does not allow predicting the capacity of agricultural or forested soil to provide soluble Si to soil porewaters where it can be used by plants. Towards better understanding of factors controlling bioavailable Si in soils, here we quantified the release rate of Si from several pairs of French forest and agricultural topsoils, developed on calcareous or loess parent material. We used mixed-flow and column reactors, in acetate and carbonate buffers and distilled water, at various pH (4–8) and time of reaction (day to month).The rate of Si release from soil (RSi) exceeded that of crystalline clay minerals by 1–2 orders of magnitude, being 5–10 times lower than that of various allophanes. The rates were weakly dependent on pH (compared to clays or phytoliths) and varied from 5 × 10−7 to 2 × 10−6 mol/gsoil/day at 4 < pH ≤ 8. In terms of Si reactivity, four studied soil groups followed the order: “calcaric cambisols ≈ hypereutric cambisols ≥ luvisols ≥ albeluvisol”. Calcaric and hypereutric cambisols as well as luvisols exhibited a weak decrease of RSi with pH increase. The rate of Si release from albeluvisol increased 2 times with a pH increase from 4 to 8. There was no measurable difference in RSi between agricultural and forest soils.The pool of labile Si in forest soils was quantified via soil column flow-through experiments. The breakthrough curves of Si demonstrated high concentrations (1.6–5.7 mg/L) over first several hours of reaction in the soil column. The water leachable Si pool ranged from 0.04 to 0.08 mg Si gsoil−1, corresponding to labile Si stock in the 0–20 cm soil of 80–160 kg ha−1. These values can meet the annual requirements of plants in forests and cultivated soils. The pool of labile Si correlated (RPearson > 0.90; p < 0.05) with Si associated with amorphous Fe and Al compounds and soluble bioavailable Si extracted using CaCl2 method, but negatively correlated with total Si content in soils and Si of phytoliths. The main pools of soil labile Si could be allophanes, organo-Fe-Al-Si compounds and adsorbed forms of Si onto Fe and Al hydroxides. We hypothesize that, because of low sensibility of RSi to type of soil and fluid pH, the majority of soils regulate Si release at some ‘universal’ rate which is further reflected in relatively narrow range of riverine Si concentrations and export fluxes across the world. Therefore, the modeling of chemical weathering and element export flux in the watersheds should incorporate the experimentally measured reactivity of the whole soil rather than individual constituting primary or secondary minerals.