In general, should I expect better image quality from a 24-70mm f/2.8 lens than from a 24-105mm f/4 lens?

Is this because people who've reviewed these lenses have higher expectations for a 24‑70mm f/2.8 lens than they do for a 24‑105mm f/4?

There's more than just a bit of something to that. There's also probably a little more going on that has to do with those who regularly use 24-105mm f/4 lenses and find them useful tend to have lower standards in terms of pixel peeping image quality than those who insist on using f/2.8 zooms.

I'm usually, but not always, in the first group. Most of what I do is distributed over the web at around 4MP or less. Though as the resolution of my FF cameras has increased from 21MP (5D Mark II) to 22MP (5D Mark III) to 30MP (5D Mark IV) over the last decade I do find myself using the 24-70/2.8 more and the 24-105/4 less than I used to do before getting a 5D Mark IV. With the 5Ds R at 50MP, you'll probably fall more into the pixel-peeping crowd than the "good enough for the web" camp.

All else being equal (which it never is) a 24-70mm f/2.8 lens should outperform a 24-105mm f/4 lens made with the same degree of commitment to design complexity and manufacturing tolerances. That's because fewer compromises must be made between 24mm and 70mm than between 24mm and 105mm. If things are optimized for 24mm, the 105mm often suffers. If things are optimized for 105mm (they rarely, if ever are), then the 24mm end often suffers. So designers walk a tightrope between the shortest and longest focal lengths of a zoom range and try to give reasonable balanced performance throughout the zoom range. At least in theory.

In practice, most wide angle to normal zoom lenses are better on the short end and not quite as good on the long end. Roger Cicala, the founder and chief lens guru at lensrental.com, calls this "Roger's Law of Wide Zoom relativity." A similarly designed 24-70mm lens would also be expected to be better at 70mm than a 24-105mm lens at 70mm, but not all of them are.

Based on my own experience, even the original EF 24-70mm f/2.8 L is noticeably better than the original EF 24-105mm f/4 L IS at common focal lengths and apertures unless camera movement when using longer exposure times is a consideration. The 24-105mm/4 has IS, the 24-70/2.8 does not. The same is true of the EF 24-70mm f/2.8 L II and the EF 24-105mm f/4 L IS II.

The latest Tamron 24-70/2.8 G2, Sigma 24-70/2.8 Art, and Sigma 24-105/4 Art all also have lens based stabilization.

Assuming your primary activities of landscape and still life photography are done from a stable tripod using best practices in terms of technique, then there may be another lens that will be just as good for your purposes.

The Canon EF 24-70mm f/4 IS is an excellent landscape lens. Similarly, unless you're just dying to get narrower f/2.8 depth of field for your still lifes, it's pretty good for that, too. It also has a near 1:1 maximum magnification of 0.70X (compared to around 0.21X to 0.25X for the others) at 70mm and minimum focus distance, so it's useful for doing macro style work. It's priced a bit less than the latest Sigma and Tamron 24-70/2.8 offerings and a tad more than the Sigma 24-105/4 ART. It's less than half the price of the EF 24-70mm f/2.8 L II.

Looking at comparisons between these lenses and the other three tested at DxO Mark, which fall in line with pretty much every other comparison I've seen of them:

At f/8, there's no real significant difference between any of these lenses.

At f/5.6 and f/4, the Sigma 24-70/2.8 ART is not quite as good at 70mm as it is at 24-50mm, nor is it as good at 70mm as the Sigma 24-105/4 or the Canon 24-70/4 are.

At f/5.6 and f/4, the Canon 24-70/4 is a little weak at 50mm¹, but is the match of all the others at 24-35mm and at 70mm. Since most of us tend to use our zoom lenses at either extreme more than we use them in the mid focal lengths, this may or may not be important for you.

The other main consideration when comparing Canon EF lenses with third party lenses is autofocus performance. Native Canon lenses tend to be faster, more accurate, and more consistent with regard to AF than third party lenses used on the same Canon camera. Sigma and Tamron do now both offer USB docks that can be attached to their lenses to update firmware and to calibrate a lens to a particular camera body. This can improve overall accuracy, but not consistency from one frame to the next. Even so, if you plan to use AF a great deal, then a native Canon lens will likely AF better than a third party lens. If you tend to manually focus your landscape and still life shots, then AF performance isn't as big a consideration for you.

_{¹ Though unexpectedly atypical, this has been noticed by pretty much everyone who has tested examples of the EF 24-70mm f/4 L IS.}