Carbon dioxide absorbers allow the use of fresh gas flow below minute ventilation (V˙E). Models are developed and tested in vitro to quantify their performance with variable carbon dioxide load (V˙CO2), fresh gas flow, V˙E, end-tidal carbon dioxide (ETco2) fraction, and the type of workstation used. First principles are used to derive a linear relationship between fresh gas flow and fractional canister usage or FCU0.5 (the reciprocal of the time for the inspiratory carbon dioxide fraction to reach 0.5%). This forms the basis for two basic models in which V˙E was measured by spirometry or calculated. These models were extended by multiplying V˙E with an empirical workstation factor. To validate the four models, two hypotheses were tested. To test whether the FCU0.5 intercept varied proportionally with V˙CO2 and was independent of V˙E, FCU was measured for 10 canisters tested with a fixed 0.3 l/min fresh gas flow and a range of V˙CO2 while V˙E was either constant or adjusted to maintain ETco2 fraction. A t test was used to compare the two groups. To confirm whether a change in V˙CO2 accompanied by a change in V˙E to maintain ETco2 fraction would shift the linear fresh gas flow-FCU0.5 relationship in a parallel manner, 19 canisters were tested with different combinations of V˙CO2 and fresh gas flow. These measured FCU values were compared to those predicted by the four models using Varvel's performance criteria. With 0.3 l/min fresh gas flow, FCU0.5 was proportional with V˙CO2 and independent of whether V˙E was adjusted to maintain ETco2 fraction or not (P = 0.962). The hypothesized parallel shift of the fresh gas flow-FCU0.5 relationship was confirmed. Both extended models are good candidate models. The models predict prepacked canister performance in vitro over the range of V˙E, fresh gas flow, and V˙CO2 likely to be encountered in routine clinical practice. In vivo validation is still needed.