I think the evidence shows that massive bodies bend light, magnetism physically cannot.
To address magnetism: photons are EM emissions, and only influenced by matter, rather than the EM field itself; this surely rules out magnetism as a bending factor. Photons, directly, are unaffected by magnetism. The electric force is matter-based, photons are just massless force carriers manifested by the electric force, only reacting through matter, not the field itself.
To address dust, nebular gas, or the like: matter and it’s force carriers can scatter light, which wouldn’t give distortion, just absorption and emission of new photons. So, dust’s mass can redirect light via space-time curvature or re-emit it via direct interaction with the electrons*. Spectral analysis can show this through matter-wavelength signatures (*to emphasize the latter: atoms will accept and re-radiate photons per their electron structure/orbit as it’s influenced by absorbed photons)
The deeper point is: light is reflected, refracted, and diffracted by interaction with matter. Gravitational lensing is redirection of photons, not by matter itself, but it’s distal action on space-time. This is observed clearly in astronomic endeavors. The most profound observations show a halo-like curve of incoming light (of a Doppler signature that is obviously sourced behind the the object). This means it’s not re-emitted by this body, but bent by it’s massive existence.
I hope a real astronomer can give an example. Otherwise, upon viewing apparent massive bodies which are specific to this phenomenon, we can see photon-curvature as well as re-emmitance. Spatial dust and other matter has a specific sign. Some light really is lensed, not re-shone. We assume the lensing is space-time curvature via mass, as there is no better explanation within the Eisteinian-Physics and QFT paradigms.
To reiterate: magnetic bodies just don’t seem to influence photons via their field, regardless of their field strength.