The performance of CoRoT in the detection of planetary transits was estimated using different approachs, from simple modeling to more elaborate tests in which detection is performed in “realistic” conditions. Some emphasis is placed on the tests made on synthetic light-curves accounting for instrumental noises, expected “stellar noises” due to stellar variability and various types of ambiguities due for example to eclipsing binaries. These tests were performed within the CoRoT community, by the various teams working on transit detection. Two collective exercices were orgenized, with a twofold aim: (i) to test CoRoT's capability to detect and identify planetary transits when placed in realistic conditions, (ii) to acustom the different teams to work in common and be prepared to a collaborative analysis of the true CoRoT data. The impact of the chromatic device is also discussed, and it is checked that in a number of cases, splitting the photometric aperture into three channels helps the discrimination of background eclipsing binaries.