Abstract. Abstract. The global NorESM1-M model that produced results for CMIP5 (http://cmip-pcmdi.llnl.gov/cmip5/index.html) has been slightly upgraded to NorESM1-Happi, and has been run with double resolution (~ 1° in the atmosphere and the land surface) to provide model simulations to address the differences between a 1.5 °C and a 2.0 °C warmer climate than the 1850 pre-industrial. As a part of the validation of temperature-targeted model simulations, the atmosphere and land models have been run fully coupled with deep ocean and sea-ice as an extension of the NorESM1-M which produced CMIP5-results. Selected results from a standard set of validation experiments are discussed: a 500-year 1850 pre-industrial control run, three runs for the historical period 1850–2005, three detection and attribution runs, and three future projection runs based on RCPs. NorESM1-Happi has a better representation of sea-ice, improved Northern Hemisphere (NH) extratropical cyclone and blocking activity, and a fair representation of the Madden-Julian oscillation. The amplitude of ENSO signals is reduced and is too small, although the frequency is improved. The strength of the AMOC is larger and probably too large. Modern era global near-surface temperatures and the cloudiness are considerably under-estimated, while the precipitation and the intensity of the hydrological cycle are over-estimated, although the atmospheric residence time of water-vapour appears satisfactory. An ensemble of AMIP-type runs with prescribed SSTs and sea-ice from observations at present-day and a set of global CMIP5 models for a 1.5 °C and a 2.0 °C world (i.e. AMIP) has been provided by the model to a multi-model project (HAPPI, http://www.happimip.org/). This paper concentrates on the results from the NorESM1-Happi AMIP runs, which are compared to results from a slab-ocean version of the model (NorESM1-HappiSO) designed to emulate the AMIP simulation allowing SST and sea-ice to respond. The paper discusses the Arctic Amplification of the global change signal. The slab-ocean results generally show stronger response than the AMIP results to a global change, such as reduced NH extratropical cyclone activity, and different changes in the occurrence of blocking. A considerable difference in the reduction of sea-ice in the Arctic between a 1.5 °C and a 2.0 °C world is simulated. Ice-free summer conditions in the Arctic is estimated to be very rare for the 1.5 °C case, but to occur 40 % of the time for the 2.0 °C case. These results agree with some fully coupled models, but need to be further confirmed.