That depends on the definition of "real 10bit". If you're simply after native 10bit panel, then I'll put out the list.
Panasonic : Only the top-end UHD Premium DX902 supports native 10bit panel. DX750 and below, they use 8bit panel.
"Unlike the top-tier DX902 however, the DX750 is not Ultra HD Premium-certified due to lack of native 10-bit panel and inability to hit 1000 nits peak brightness" (HDTVTest.co.uk)
LG : The entire UHD lineup. UH6100 and above.
"The color gradient is pretty good for an 8 bit panel. There are not any banding issues that we can see from our gradient pattern. The only little imperfections that you can see when looking at the picture are more related to the gray uniformity.
Update 09/30/2016: Our original test was showing an incorrect color depth of 8 bit due to some incorrect drivers on our system, but after some correction to our test apparatus, we tested again the color depth and we can confirm that the LG UH6100 does in fact have an 10 bit panel." (RTings)
Sony : The entire UHD lineup. X700D and above.
"The Sony X700D has a good 10 bit panel and the color reproduction is pretty even. There is not really any obvious banding problem that can be seen on our test picture that could cause a problem when watching movies." (RTings)
Samsung : The entire UHD lineup. KU6300 series and above
"The Samsung KU6300 can display our gradient test image fairly well. On our test picture, the gradation is smooth overall in the light shades with some small anomalies in the darker shades, especially in the green color. But it should not be an issue in regular content.
Update 10/26/2016: Our original test was showing 8 bit gradations due to incorrect drivers on our system. After some correction to our test apparatus, we have retested the color depth and found that it is able to display a 10 bit gradient smoothly."
However, I would take RTings interpretation with a huge grain of salt. They say it's due to driver error, but their Samsung KS series reviews at that time still were able to resolve 10bits. PQ EOTF is efficient that 8bit + FRC display can still pack up to 400 nits without banding. Here's HDTVTest's results.
"Although greyscale was flattened after calibration, minor banding and discolouration remained visible on a grey ramp pattern, suggesting that the underlying panel wasn’t true 10-bit." (HDTVTest LG UH7700 review)
So, I'm more inclined to believe only Samsung's SUHD lineups are true 10bits, and for LGs, only the OLED TVs. I'm more inclined to believe Sony's results though as for last years models, they were the only ones along with Panasonic that were able to resolve true 10bit signal when fed a signal generator. The Samsung JS SUHD series and 2015 LG OLED TVs, despite having true 10bit panel were only able to resolve 8bit signal.
That doesn't mean Samsung KS series are home free either. Because they've decided on 4000 nits, tone mapping system, compression was so severe, posterization has appeared in a form of banding.
"Unfortunately in an effort to correctly track PQ (perceptual quantizer) EOTF (electro-optical transfer function) whilst resolving detail up to 4000+ nits, Samsung’s tone-mapping algorithm introduced visible posterisation in what should be smooth gradients (e.g. the skies in The Revenant and The Martian). These banding artefacts could be reduced but not completely eradicated by switching to the [Standard] picture preset or engaging [HDR+ Mode], but once we measured and calibrated these modes to track PQ curve and Rec.2020 colours as close as possible (yes, we leave no stone unturned), the posterization reared its ugly head again. Of course, [HDR+ Mode] involved other compromises too (as we’ll describe in the next section), making it a no go." (HDTVTest Samsung KS9500/KS9800 review)
I wouldn't worry too much about true 10bit panel for entry level/mid range TVs though. Most of them have brightness equal to or less than 400 cd/m2, so even when using 8bit panel with dithering, banding shouldn't be too much of an issue, as RTings has shown.
