You're comparing primes vs zooms. This initial test suggests that the new lens is sharper across the frame than the current class-leading zooms (Canon 100-400, Sigma 120-300, Canon 200-400). It also matches some of the primes at comparable apertures - obviously not the very best ones (e.g. 400 f/2.8, of either Canon or Nikon make) but a significant number of the others. Will need to wait for proper MTF charts to be published (most likely by Lensrentals) but it's looking promising.
There's no comparison between the utility of zooms and primes when shooting landscapes. 'Foot-zooming' doesn't work - it changes the relative size and position of different scene elements within the frame, or requires you to move to an inaccessible location (off a cliff, into the air, etc.). Cropping costs you valuable megapixels. So, in general, the focal length you need is dictated by the shooting position you need (or the shooting position you can get) and is rarely a round number. A 300mm lens does you no good if you actually need 270mm or 370mm for a proper composition - it's too long for the former, while you're cropping for the latter. A good zoom set to 370mm will give you more detail than a 300mm prime cropped to the same angle of view, particularly if both lenses are stopped down (as is typical for shooting landscapes).
When shooting wildlife or sports, you often have to zoom back and forth a lot, especially in situations where subjects are large or can come very close (e.g. shooting animals in East/South Africa, or trackside at an athletics competition). One moment you might need 500mm, the next you need 200mm or shorter. No time to change lenses here. That's why a combination of 200-400 with inbuilt 1.4x TC on one body and a 500 f/4 (with 1.4x TC attached) on another body is such a potent combination for a dedicated wildlife trip. You've got the zoom for anything that's large, close or both, as well as the prime for small or distant subjects where you're focal length constrained and just need as many pixels to crop from as possible.
As for your specific examples, look at the 'Profiles' under the 'Sharpness' tab. This gets rid of all the garbage about 'perceptive megapixels', the 'transmission' value that gives a bonus to fast lenses and a penalty to slow ones, etc. and just looks purely at the sharpness.
At 300mm f/5.6, the Sony is a bit sharper than the prime in the centre and a bit weaker in the edges (and 300mm seems to be the Sony's weak spot - it's better at both 200mm and 400mm). If you're shooting at 350mm, vs cropping from the 300mm prime, you've more than lost this advantage at the edges.
With the Sony at 100mm f/4.5 and the Zeiss at 135mm f/4, the Sony is a bit sharper in the centre, and just as sharp at the edges (although the Sony has more astigmatism there - the tangential sharpness is better than the prime, but the sagittal is worse)
The primes do better distortion-wise, but none of them have high distortion, and low to moderate distortion barely matters when shooting landscapes (not too many straight lines), especially at long focal lengths.
Also, why would you compare 300 prime plus TC with 100-400 plus TC? The 300 prime will reach 420mm with a 1.4x TC. The 400mm can already reach 400mm without a TC. So it makes more sense to compare the 300mm with 1.4x TC vs 100-400 without TC, or 300mm with 2x TC vs 100-400 with 1.4x TC.
