Context : Among the myriad of data collected by the ESA Gaia satellite , about 150 million spectra will be delivered by the Radial Velocity Spectrometer ( RVS ) for stars as faint as G _ { RVS } \sim 16 .
A specific stellar parametrization will be performed for most of these RVS spectra , i.e .
those with enough high signal-to-noise ratio that should correspond to single stars having a magnitude in the RVS band brighter than \sim 14.5 .
Some individual chemical abundances will also be estimated for the brightest targets .

Aims : We describe the different parametrization codes that have been specifically developed or adapted for RVS spectra within the GSP-Spec working group of the analysis consortium .
The tested codes are based on optimization ( FERRE and GAUGUIN ) , projection ( MATISSE ) or pattern recognition methods ( Artificial Neural Networks ) .
We present and discuss their expected performances in the recovered stellar atmospheric parameters ( effective temperature , surface gravity , overall metallicity ) for B- to K- type stars .
The performances for the determinations of [ \alpha /Fe ] ratios are also presented for cool stars .

Methods : The different codes have been homogeneously tested with a large grid of RVS simulated synthetic spectra of BAFGK-spectral types ( dwarfs and giants ) with metallicities varying from 10 ^ { -2.5 } to 10 ^ { +0.5 } the solar metallicity , and considering variations of \pm 0.4 dex in the composition of the \alpha -elements .
The tests have been performed for S/N ratios ranging from 10 to 350

Results : For all the considered stellar types , stars brighter than G _ { RVS } \sim 12.5 will be very efficiently parametrized by the GSP-Spec pipeline , including solid estimations of [ \alpha /Fe ] .
Typical internal errors for FGK metal-rich and metal-intermediate stars are around 40 K in T _ { eff } , 0.10 dex in log ( g ) , 0.04 dex in [ M/H ] , and 0.03 dex in [ \alpha /Fe ] at G _ { RVS } =10.3 .
They degrade to 155 K in T _ { eff } , 0.15 dex in log ( g ) , 0.10 dex in [ M/H ] , and 0.1 dex in [ \alpha /Fe ] at G _ { RVS } \sim 12 .
Similar accuracies in T _ { eff } and [ M/H ] are found for A-type stars , while the log ( g ) derivation is more accurate ( errors of 0.07 and 0.12 dex at G _ { RVS } =12.6 and 13.4 , respectively ) .
For the faintest stars , with G _ { RVS } \ga 13-14 , a T _ { eff } input from the spectrophotometric derived parameters will allow the improvement of the final GSP-Spec parametrization .

Conclusions : The reported results , although neglecting possible mismatches between synthetic and real spectra , show that the contribution of the RVS based stellar parameters will be unique in the brighter part of the Gaia survey allowing crucial age estimations , and accurate chemical abundances .
This will constitute a unique and precious sample for which many pieces of the Milky Way history puzzle will be available , with unprecedented precision and statistical relevance .