As woodgeek said, this is a complicated problem.
i) The rate of accumulation of biomass.
ii) That depends upon both the rate of photosynthesis, which is influenced by the species, and the local conditions, but also,
iii) The countervailing respiration that releases CO2, both directly from the plant, and also from the dead plant material.
iv) Respiratory breakdown of the dead material depends upon both the chemical composition of the material (lettuce leaves decompose faster than oak logs), but also the local conditions (leafs in North Carolina break down much faster than leaves in Labrador).
v) The breakdown of the dead material determines the accumulation of carbon in the soil, which as woodgeek wrote, can be substantial.
In terms of management, there is also the issue of the non-carbon inputs into biomass.
One of the attractions of Switchgrass/Miscanthus as a possible biofuel or cellulose course is that the harvested mature grass contains little other than fixed carbon, so it does not drawn down the local soil nutrient stocks much.
Brazilian sugarcane with nitrogen fixing symbionts is also good in that regard, in comparison to a heavy feeder like USA corn.
In my area some of the soils are so poor that after 3-4 rounds of top-grading and then clear cutting, there may not be enough nutrient left in the soil to regenerate a full forest canopy. The nutrients were carried off in the wood.
On top of all this, is the risk of releasing methane.
Carbon for carbon, methane has about 24X the warming potential of CO2. So if 5% of the plant material is converted to methane that is released, that cancels the uptake of CO2 into the rest of the biomass.
But in general, growing something, particularly perennials, is better than growing nothing.