The Community Climate System Model Version 4 (CCSM4) was completed and released in 2010. This paper describes the improvements in all CCSM components compared to the previous version, CCSM3. CCSM4 produces more realistic El Niño–Southern Oscillation (ENSO) variability with a more realistic frequency distribution, although the amplitude is too large compared to observations. A new overflow parameterization in the ocean component improves the simulation of the Gulf Stream path and the North Atlantic Ocean meridional overturning circulation. Changes to the land component lead to a much improved annual cycle of water storage, especially in the tropics. The CCSM4 sea ice component uses more realistic albedos and improves the Arctic sea ice concentration. An ensemble of twentieth-century simulations produces a good match to the observed September Arctic sea ice extent from 1979 to 2005. The CCSM4 ensemble mean increase in globally averaged surface temperature between 1850 and 2005 is larger than the observed increase by about 0.4°C. This is consistent with the fact that CCSM4 does not include a representation of the indirect effects of aerosols. CCSM4 still has significant biases, such as the mean precipitation distribution in the tropical Pacific Ocean, too much low cloud in the Arctic, and the latitudinal distributions of shortwave and longwave cloud forcings. CCSM4 has a completely new coupling infrastructure, CPL7, which allows for greater flexibility in running model components. The CCSM4 preindustrial control and twentieth-century runs show improvements in sea surface temperature, ocean meridional overturning circulation, Arctic sea ice concentration, tropical precipitation, and Madden–Julian oscillation. However, CCSM4 still has some biases, such as the mean precipitation distribution in the tropical Pacific Ocean, too much low cloud in the Arctic, and the latitudinal distributions of shortwave and longwave cloud forcings. The CCSM4 has a better representation of the tropical Pacific climatology and variability, including improved precipitation, Madden–Julian oscillation, and El Niño–Southern Oscillation variability. The CCSM4 also shows improvements in land climatology and variability, including better representation of land water storage and land surface temperatures. The CCSM4 has a more realistic simulation of the tropical Pacific climatology and variability, including improved precipitation, Madden–Julian oscillation, and El Niño–Southern Oscillation variability. The CCSM4 has a better representation of the land climatology and variability, including better representation of land water storage and land surface temperatures. The CCSM4 has a more realistic simulation of the climate evolution of twentieth-century runs, including improved globally averaged surface temperature and other climate variables.The Community Climate System Model Version 4 (CCSM4) was completed and released in 2010. This paper describes the improvements in all CCSM components compared to the previous version, CCSM3. CCSM4 produces more realistic El Niño–Southern Oscillation (ENSO) variability with a more realistic frequency distribution, although the amplitude is too large compared to observations. A new overflow parameterization in the ocean component improves the simulation of the Gulf Stream path and the North Atlantic Ocean meridional overturning circulation. Changes to the land component lead to a much improved annual cycle of water storage, especially in the tropics. The CCSM4 sea ice component uses more realistic albedos and improves the Arctic sea ice concentration. An ensemble of twentieth-century simulations produces a good match to the observed September Arctic sea ice extent from 1979 to 2005. The CCSM4 ensemble mean increase in globally averaged surface temperature between 1850 and 2005 is larger than the observed increase by about 0.4°C. This is consistent with the fact that CCSM4 does not include a representation of the indirect effects of aerosols. CCSM4 still has significant biases, such as the mean precipitation distribution in the tropical Pacific Ocean, too much low cloud in the Arctic, and the latitudinal distributions of shortwave and longwave cloud forcings. CCSM4 has a completely new coupling infrastructure, CPL7, which allows for greater flexibility in running model components. The CCSM4 preindustrial control and twentieth-century runs show improvements in sea surface temperature, ocean meridional overturning circulation, Arctic sea ice concentration, tropical precipitation, and Madden–Julian oscillation. However, CCSM4 still has some biases, such as the mean precipitation distribution in the tropical Pacific Ocean, too much low cloud in the Arctic, and the latitudinal distributions of shortwave and longwave cloud forcings. The CCSM4 has a better representation of the tropical Pacific climatology and variability, including improved precipitation, Madden–Julian oscillation, and El Niño–Southern Oscillation variability. The CCSM4 also shows improvements in land climatology and variability, including better representation of land water storage and land surface temperatures. The CCSM4 has a more realistic simulation of the tropical Pacific climatology and variability, including improved precipitation, Madden–Julian oscillation, and El Niño–Southern Oscillation variability. The CCSM4 has a better representation of the land climatology and variability, including better representation of land water storage and land surface temperatures. The CCSM4 has a more realistic simulation of the climate evolution of twentieth-century runs, including improved globally averaged surface temperature and other climate variables.