Health benefits of probiotics in humans essentially depend on their ability to survive during gastrointestinal (GI) transit and to modulate gut microbiota. To date, there is few data on the impact of galenic formulations of probiotics on these parameters. Even if clinical studies remain the gold standard to evaluate the efficacy of galenic forms, they stay hampered by technical, ethical and cost reasons. As an alternative approach, we used two complementary in vitro models of the human gut, the TNO gastrointestinal (TIM-1) model and the Artificial Colon (ARCOL), to study the effect of three oral formulations of a Lactobacillus salivarius strain (powder, capsule and sustained-release tablet) on its viability and interactions with gut microbiota. In the TIM-1 stomach, no or low numbers of bacteria were respectively released from the capsule and tablet, confirming their gastro-resistance. The capsule was disintegrated in the jejunum on average 76 min after administration while the core of sustained-release tablet was still intact at the end of digestion. Viability in TIM-1 was significantly influenced by the galenic form with survival percentages of 0.003±0.004%, 2.8±0.6% and 17.0±1.8% (n=3) for powder, capsule and tablet, respectively. In the ARCOL, the survival of the strain tended to be higher in the post-treatment phase with the tablet compared to capsule, but gut microbiota composition and activity were not differently modulated by the two formulations. In conclusion, the sustained-release tablet emerged as the formulation that most effectively preserved viability of the tested strain during GI passage. This study highlights the usefulness of in vitro gut models for the pre-screening of probiotic pharmaceutical forms. Their use could also easily be extended to the evaluation of the effects of food matrices and age on probiotic survival and activity during GI transit.