Critical power derived from the 3-min all-out test (3MT) was recently utilized to estimate the exercise intensity boundaries in competitive cyclists. Considering that physiological testing is challenging in swimming, the purpose of this study was to examine whether critical speed (CS) derived from 3MT could be used for the same purpose in swimming. The second aim was to assess the accuracy of the 50-40 and 30-20 beats below maximal heart rate method (BBM), currently utilized by swimming coaches to demarcate boundaries between moderate-heavy and heavy-severe exercise, respectively. Thirteen swimmers completed an incremental step test (IST) and 3MT in freestyle to establish speeds at: lactate threshold (LT); lactate turnpoint (LTP); maximum aerobic speed (SMAX); and CS. Using linear regression through origin, speeds at LT, LTP and SMAX were predicted at 89, 98 and 104% of CS derived from 3MT. There were no significant differences between threshold speeds derived from IST and 3MT (p>0.05), and nearly perfect correlations at LT (1.21±0.06; 1.21±0.06 m.s-1; r=0.92) and LTP (1.33±0.07; 1.33±0.07 m.s-1; r=0.90), and very large correlations at SMAX (1.40±0.06; 1.40±0.07 m.s-1; r=0.88; all p<0.0001). Speeds estimated at 50 (1.11±0.08 m.s-1) and 40BBM (1.17±0.07 m.s-1) were lower compared to LT, and speeds estimated at 30 (1.23±0.07 m.s-1) and 20BBM (1.29±0.07 m.s-1) were lower compared to LTP and CS (all p<0.02). The 3MT can therefore be used as an alternative to IST to estimate exercise intensity boundaries, in practical settings where resources or time might be limited. However, the BBM significantly underestimates speeds at LT, LTP and CS.