Revenue from carbon credits from rainforest stands could encourage reforestation for biodiversity conservation on private land in north-eastern Queensland, Australia. Current models and allometrics for estimating carbon, however, are not calibrated against sites in the region and underestimate carbon stocks. We assess the accuracy of the two accepted methods to estimate carbon stocks in Australian rainforests: FullCAM and the Keith et al. (2000) allometric. We also assess the effect of FullCAM's discounting of small stems (2.5-10cm) to carbon stocks, and compare the carbon benefits of the three reforestation methods in the region to identify planting configurations with the best carbon sequestration potential. We sampled 27 rainforest stands in north-eastern Queensland. Using these data we calculated above-ground biomass (AGB) using the Keith allometric and derived the above-ground carbon (AGC). We compared our estimates across three reforestation methods with the FullCAM modelled estimates for the same sites, and with estimates derived from two global rainforest allometrics (Brown, 1997; Chave et al., 2005). The Keith allometric estimated that planted forests yielded on average 20Mg of tradable carbon ha -1y -1 (i.e. CO 2-equivalent), with no differences between plantation forests and environmental plantings, although the former had more large diameter stems. Small stems (<10cm) accounted for 15.1% of AGB in plantings <20years old. However, even excluding these, the estimates using the Keith allometric were 19.5% greater than those of FullCAM; the Chave allometric 40.4% greater; and the Brown allometric 54.9% greater. More thorough forest mensuration using actual tree volumes and densities is required to determine a biomass allometric function for rainforests in the region. Until then, we recommend the Chave allometric function. It provides intermediate values, is based on the widest range of tropical trees and has been shown to be accurate away from the sites used for its development. This study demonstrates the inadequacy of current methods for estimating carbon stocks in rainforest plantings in north-eastern Queensland. A tailored allometric and the re-parameterisation of FullCAM is needed to reflect both the region's environmental characteristics and the vegetation structure of young reforestation stands. Current estimates deprive landholders of financial incentives and underestimate the national greenhouse gas benefits of tree planting in the Wet Tropics.