The NNPDF Collaboration presents the first determination of parton distributions (PDFs) of the nucleon at next-to-leading order (NLO) and next-to-next-to-leading order (NNLO) based on a global data set that includes LHC data: NNPDF2.3. The data set includes deep inelastic scattering, Drell-Yan, gauge boson production, and jet data, as well as all relevant LHC data with available experimental systematic uncertainties. The collaboration introduces an improved implementation of the FastKernel method and a more effective minimization methodology. The NNPDF2.3 PDF sets are compared to the NNPDF2.1 sets to assess the impact of the LHC data, finding broad consistency among all LHC data and with older datasets. Predictions for various standard candle cross-sections are presented and compared to those obtained using NNPDF2.1, with a specific focus on the impact of ATLAS electroweak data on the determination of the strangeness fraction of the proton. The paper also discusses collider PDF sets constructed using only data from HERA, Tevatron, and LHC, but finds that this dataset is neither precise nor complete enough for a competitive PDF determination.The NNPDF Collaboration presents the first determination of parton distributions (PDFs) of the nucleon at next-to-leading order (NLO) and next-to-next-to-leading order (NNLO) based on a global data set that includes LHC data: NNPDF2.3. The data set includes deep inelastic scattering, Drell-Yan, gauge boson production, and jet data, as well as all relevant LHC data with available experimental systematic uncertainties. The collaboration introduces an improved implementation of the FastKernel method and a more effective minimization methodology. The NNPDF2.3 PDF sets are compared to the NNPDF2.1 sets to assess the impact of the LHC data, finding broad consistency among all LHC data and with older datasets. Predictions for various standard candle cross-sections are presented and compared to those obtained using NNPDF2.1, with a specific focus on the impact of ATLAS electroweak data on the determination of the strangeness fraction of the proton. The paper also discusses collider PDF sets constructed using only data from HERA, Tevatron, and LHC, but finds that this dataset is neither precise nor complete enough for a competitive PDF determination.