As of the time of writing this is what the HYCOM thickness model looks like
There certainly does seem to be an improvement since last year. However this must be taken in the context of previous years too. It is possible for one year to buck the trend because of favorable meteorology, and it is widely thought that this is at least partly the reason the ice has seemed to have fared significantly better this year.
The PIOMAS model does illustrate said trend nicely. Notice that 2013 is above 2012, the same effect can be seen on the overall volume of the arctic sea ice by PIOMAS.
The trend can also be seen quite nicely on the following ice age animation. Note ice age is a useful proxy for thickness. Older ice tends to be thicker and more resistant to summer melt as a general rule. Ice that is 2 years old or more is commonly called multi-year ice (MYI).
(source: ftp://ccar.colorado.edu/pub/tschudi/iceage/gifs/ ; the ice age maps were imported from the university of colorado and this GIF amination was created using the most recent week (31) of each year since 1983).
It may seem more useful to see the whole series than just compare over two years for instance 2012 and 2013 which are seen below on the left and right respectivly.
Note: Week 31 is the week commencing August 12th, this was the most recent 2013 map available. So it is not appropriate to directly compare this with the HYCOM thickness model at the top.
Actually I did something similar with the PIPS 2.0 model which I believe was a forerunner to the HYCOM on my YouTube channel and shows the 'noughty' years when it was in opperation (note the HYCOM does not show show archives further back than 2012).
(source http://www7320.nrlssc.navy.mil/pips2/ - images are no longer available, this avi was compiled for September 15th when it was in operation)
To try and give as a complete a picture as possible. I will also include a comparision of 2013 with 2012 on the TOPAZ4 model. Unfortunately TOPAZ4 only goes back to autumn 2011 so I cannot do a longer animation.
The image on the left is 2012, and the image on the right is 2013. I managed to get this for the 3rd September on both years so a direct comparison with the HYCOM model is perhaps more appropriate. Interestingly 2012 seems to have the more abundant thicker ice in the Canadian archipelago region in particular.
Finally it is worth looking at the observations we have of the thickness. I have written the thicknesses of all the current buoy data (on their map provided) in the approximate (I will redo this more accurately using identical pixel colours if necessary) HYCOM thickness colour code (but note: be very wary of a direct comparison, individual floes may vary substantially in thickness). However any talk about thickness would be a little empty without these. I have also recopied the current 2013 HYCOM image on the right, so people don't have to rely on my approximate colour matching skills.
In case the numbers arn't legible then bottom left to top right they read: (1.57, 1.32, 2.67, 2.16, 2.67, 1.39, 1.39, 1.69, 1.21).
It does seem that the HYCOM model may be over egging the thickness slightly, although the distribution seems very good. However I don't think there is enough data to make this claim with any great confidence.
Anyway my point of all this was, it seems better to try and look at a trend if possible using all the thickness data available (and to try and keep posts as informative as possible). There is an archive for past buoy data however it cannot be used for any meaningful comparison, at least not by any crude methodology that I can employ due to the large spatial differences between each year.