0.065 ± 0.040 W m−2 applied over the Earth's surface area.
You can read the abstract here, but deep ocean heating is only about 1/10th of total ocean heating. I guess we'll know more when deep Argo comes out, which I suspect will reduce the error bars significantly. Not sure when that will be, though.
Remember, it's only been since 2005 that Argo has been measuring the temperature and salinity of the top half (< 2000 m) of the ocean. If deep Argo happens by, say, 2020, that will be an amazing achievement -- because ocean heat content is by far the best way to measure the global energy imbalance that causes global warming.