Final Activity Report Summary - CIFIST (Cosmological Impact of the First Stars) The CIFIST team has been financed by a Marie Curie Excellence Grant, in the framework of the FP6 programme, in the period 2005--2009 and hosted by GEPI-Observatoire de Paris. Six researchers of different levels were supported, for different time intervals to work on the project. The goals were to provide a significant increase in the knowledge of the chemical composition of the most ancient stars in the Galaxy and to contribute to the discovery of more of these rare objects. Both goals have been attained and the results published in refereed journals. Among the most striking results are those resulting from the investigation of the lithium abundances in metal-poor stars. Lithium is one of the few nuclei which has been produced in the primordial Universe and its primordial abundance is an indirect measure of the Universal baryonic density. For a long time it has been considered that such abundance coincides with the constant lithium abundance observed in metal-poor un-evolved stars (the so-called Spite plateau). The CIFIST Team has discovered an excess dispersion in the Li abundances at the lowest metallicities (below 1/1000 the solar metallicity), indicating the existence of a physical phenomenon which causes such a dispersion. At the same time the Team conducted several studies of Li in Globular Clusters, arriving at the first measurement of Li in the Main Sequence (MS) stars of a Globular Cluster and detecting a difference in Li content between MS stars and subgiants. Again this implies the existence of phenomena which alter the photospheric Li abundance. For the discovery of new and extremely metal-poor (EMP) stars the CIFIST Team developed a method to analyse the low resolution spectra of the Sloan Digital Sky Survey (SDSS) and select EMP stars. The high resolution follow-up observations confirmed the soundness of the method. Twenty such stars have already been observed, more observing proposals have been submitted to observe others. The chemical analysis of the C-rich stars (three objects) has already been accepted for publication and the analysis of the others will shortly follow. However CIFIST was highly successful beyond the initial goals. The team conducted the computation of a grid of 3D hydrodynamical simulations of stellar atmospheres (3D models for short), which is the largest available to date (about 70 models). These models are more physically motivated than classical hydrostatic 1D model atmospheres which make the chemical analysis based on these models more reliable. In this respect CIFIST placed itself on the cutting edge of spectroscopic research, being one of the two teams in the world to possess the know-how and capability of performing such computations. In order to validate the 3D models, CIFIST undertook an extensive analysis of the solar chemical composition, completing the study of about 10 chemical elements, among which the highly significant C,N,O and Fe. The CIFIST Team was able to show, that the use of 3D models does not lead to lower solar abundances, contrary to previous claims. On the topic of solar photospheric abundances, the CIFIST became one of the two leading groups in the world.