Just today the BICEP/KECK group just placed a new manuscript on arxiv with their latest data on BB modes. Included in the new paper is data collected by the KECK telescope at 95 GHz. (The new data is in red in the figure below.)
The gist of the paper is that the new data at 95 GHz finds only very slight (if any) evidence for BB-modes from tensor-modes (i.e. inflationary-based gravitational waves.) Right now, there's not enough data to rule out dust. The error bars are just too large to say anything conclusively. The point here is that more data at 95 GHz (where there is less signal from dust than at 150 GHz) does not help to confirm the tensor-mode nature of the signal found at 150 GHz. (Note that this does not help or hurt inflation theories in general; it only helps to rule out a few of the theories that predict large values of r. In my opinion, slow-roll inflation with the Higgs field as the slow-roll scalar field for inflation is a pretty strong theory. The question is: what is the scale of inflation? And this new data seems to suggest that the energy scale for inflation is low enough that it will be difficult to experimentally measure tensor-mode gravitational waves over the next decade.)
When you couple the data above with the lack of meaningful BB data from Planck at low multipole l (see figure below), you are left with the following conclusions:
The ratio of initial tensor-to-scalar power must be small (i.e. r < 0.07.) And assuming that the tilt of the tensor power (i.e. nT) is small and negative in value (as predicted by slow-roll inflation), then the ratio of tensor-to-scalar power, r, must be really small (likely less than 0.05.) The Planck BB data between l =2 and l=16 looks like junk. The is no discernible peak around l = 4 and minimum around l =9 (as supposedly is found in the TE and EE data. Though, the TE and EE data looks like junk to me.) Basically, the data below is evidence that both tensor power and the reionization optical depth are quite small compared to the sensitivity of the Planck telescope.
To conclude, I'd like to show some interesting results from Hunt and Sarkar 2015, who used data from multiple group to try to determine the initial, inflationary-induced scalar and tensor power spectra.
Since Planck found that there is no evidence for BB modes at low multipole, but BICEP/KECK data (before the new 95 GHz data) is compatible with a tensor mode (if dust is small), Hunt and Sarkar 2015 tried to make tensor power curves that fit both data sets. However, note also that in this case, the tensor power tilt (i.e. nT) would be large and positive (and this is not favored by current slow-roll inflationary theories.) Therefore, Hunt and Sarkar 2015 concluded that the BICEP/KECK data was likely due to dust simply because it would imply a positive tensor power tilt.
Therefore, my conclusion is the same as Planck 2015 and Hunt and Sarkar 2015: the value of the initial tensor-to-scale ratio is likely less than 0.05 and we can model inflation (and hence understand why there are any perturbations in our universe) using a slow-roll scalar potential (such as the Higgs fields.)