I'm not so confident.
The highest speed, with shortest copper, rarely gets even close to 1Gbps when the 2-23MHz spectrum is being excluded. If the orange line started around 800Mbps, I'd believe it.
And, as the work is done by a German institute, wouldn't we expect them to perform work one a cable model closer to home? 0.4mm seems common in Europe, though I think I've seen 0.6mm is common in Switzerland. Austria is also home to a fair amount of G.Fast interest.
There are lots of conditions that apply to any g.fast graphs, and none of them are specified here. I would be happier with a graph like this one from Swisscom in 2015, with all the attendant details:
I'd love it if we could get graphs that show the new results (increased bit, power and PSD) with the same level of detail. Especially if it came with an idea of the cable model:
It is interesting to see from that model that (in Switzerland, at least) the 0.5mm "standard" copper has less attenuation than 0.6mm copper which is insulated by paper. It just goes to show that the copper alone isn't the only factor!
I *think* the reference "CAD55" cable is actually one of BT's cables - a 4-pair dropwire, with 0.5mm copper. If so, then BT's results are likely to be (currently) just below the blue dots on the Pe4D graph, but better than the Paper graph.