1409 has a sort of watercolor effect, meaning smeary detail at the pixel level when zoomed in and examined; it looks like maybe a good deal of noise reduction was applied, possibly due to a fairly high ISO setting. Same with the first squirrel shot. The bird photo shows a lot of chroma noise in the smooth areas of the image, but the bird appears well-focused and sharp-ish. Honestly, it looks to me that the weaker link is sensor performance on these. I cannot get any EXIF info, and just looked at these magnified to 2x on-screen.
If any image was grossly under-exposed, and then processed, it's not a good, fair way to evaluate the image quality the lens can produce. Shot 1409 for example, that is the classic "watercolor effect" of noise reduction...the per-pixel info sort of breaks apart when you zoom in...it's not necessarily the LENS that is doing the damage.
As far as pictures coming out darker than expected: I would look closely at the camera settings (accidental Minus Exposure comp for example), and also consider that with a long telephoto, many times the metering will be measuring a VERY small area of the real world, as opposed to using a wider-angle lens that has a much broader area of acceptance. With darker subjects in the metering area, you might tend to get dimmer exposures. ALSO, with the HUGE number of lens elements in a super-zoom, many times the actual T-stop or actual transmission of light, is MUCH lower than the computed f/stop value. EACH air-to-glass surface loses light, even when multi-coated, and superzooms with 19,20,21,23 elements can have lowish T-stops compared to say, a 7-element prime lens.