This means about 340 lb-ft of torque calculated for the motor output based on measurement at the hubs and likely an assumption of perfect gearbox efficiency. The calculation will be correct, except for error due to that false assumption, as long as they entered the right gear ratio (9:1 according to the owner's manual).... the Model 3 with an 80% charge put down 340whp and about the same amount of torque. The X-Axis displays the motor RPM, and torque displayed is torque at the motor, not at the wheels.
It's too bad they did all of this setup and just did one accelerating run, starting from way up at 2000 rpm at the motor and not under full load until way above that (which makes sense for an engine but not an electric motor). The portion of the displayed curves below 4000 rpm is meaningless. A Dynapack is capable of holding any desired speed to see if the car can sustain a load, and can accelerate or decelerate.... who knows, maybe it has a minimum speed target, but it should have at least run at full load from that minimum, right?
The usable portion of the curves shows the expected behaviour. The peak-power speed (top of the constant-torque range, beginning of the constant or gradually declining power range) appears to be at about 5200 rpm. The Model 3 owners manual doesn't provide motor specifications, but for comparison the Model S manual shows the peak power point for the small motor to be at 6100 rpm. It's not surprising that this characteristic speed is lower for the Model 3's PM AC motor; I'm a bit surprised that it is still so high.
For anyone puzzled by the idea that the Model 3 is new, since some have been in customer's hands in the U.S. for a while, I'll note that yesterday was the first day of deliveries in Canada. For some reason Tesla has been stacking up piles of the things, and held them back from buyers until the big day.