During the past decade an increasing number of studies have used different fish species to study cognition. Operant conditioning is usually investigated on single individuals placed in diverse devices such as shuttle box, t-maze or plus maze, using either simple response discrimination (right vs left) or a visual cue (e.g. red card) marking the correct choice. The main concern is that learning is likely to be impaired by acute stress responses induced by fish handling often required between each trial and isolated conditions. Here we investigate for the very first time the operant conditioning flexibility of several groups of sea bass (Dicentrarchus labrax) using two self-feeder devices that were alternately activated. We also tested the impact of the positive appetitive reinforcement duration reduction over time. Fish were tested under group conditions but with a design allowing to follow them individually without any stress induced by handling and without any limitation of the number of learning trials and sessions. Six tanks of 50 PIT-tagged fish were equipped with two computerized self-feeders systems (with a PIT Tag detection antenna) in each tank. The two self-feeders in each tank were alternatively activated in order to assess the ability of the fish to identify the operational device. Triggering activity was recorded continuously for 117 days. The results showed that fish learned the task very quickly and the duration of the reinforcement period (from 7 days to 1 days) did not impair the learning process. We also showed that the high triggering fish i.e. one or two animals responsible for most of the triggering activity were the same on both self-feeders in each tank. These results open up new opportunities to investigate fish learning and cognition in response to the main drawbacks of experiments performed in maze as highlighted above.