The standard model of cosmology is based on Einstein’s theory of general relativity. In order to account for cosmological observations this has required the introduction of dark matter – which also helps explain the properties of individual galaxies – and dark energy. The result model, which I would describe as a working hypothesis, is rather successful but it is reasonable to question whether either or both of the dark components can be avoided by adopting an alternative theory of gravity instead of Einstein’s.
There is an interesting paper by Kris Pardo and David Spergel on arXiv that argues that none of the modifications of Einstein’s theory currently on the market is able to eliminate the need for dark matter. Here is the abstract of this paper:

It’s a more sophisticated version of an argument that has been going around at least in qualitative form for some time. The gist of it is that the distinctive pattern of fluctuations in the cosmic microwave background, observed by e.g. the Planck experiment, arise from coupling between baryons and photons in the early Universe. Similar features can be observed in the distribution of galaxies – where they are called Baryon Acoustic Oscilations (BAO) at a more recent cosmic epoch, but they are are much weaker. This is easily explicable if there is a dark matter component that dominates gravitational instability at late times but does not couple to photons via electromagnetic interactions. This is summed up in the following graphic (which I think I stole from a talk by John Peacock) based on data from about 20 years ago:

If there were no dark matter the coherent features seen in the power spectrum of the galaxy distribution would be much stronger; with dark matter dominating they are masked by the general growth of the collisionless component so their relative amplitude decreases.
The graphic shows how increasing the dark matter component from 0.1 to 0.3, while keeping the baryon component fixed, suppresses the wiggles corresponding to BAOs. The data suggest a dark matter contribution at the upper end of that range, consistent with the standard cosmology.
Of course if there are were no baryons at all there wouldn’t be fluctuations in either the CMB polarization or the galaxy distribution so both spectra would be smooth as shown in the graphic, but in that case there wouldn’t be anyone around to write about them as people are made of baryons.
This general conclusion is confirmed by the Pardo & Spergel paper, though it must be said that the argument doesn’t mean that modified gravity is impossible. It’s just that it seems nobody has yet thought of a specific model that satisfies all the constraints. That may change.