The configuration is about as good as it gets for going ductless- most of the gain in the bedroom end is before the mid-day heat, and the place that's getting a mid-afternoon roasting is the open area where you'd most likely put the mini-split head. The question is whether you can keep the bedroom cool enough for sleeping via convection alone.
Even insulated ducts in the attic above the insulation are likely to be adding at least ~10,000BTU/hr to the peak load, plus infiltration air leakage from room-to-room pressure differences generated by the air-handler may be adding a bit more. Rule of thumb "typical" losses related to sealed-insulated ducts outside of conditioned space would be ~20% of total range, but it varies a lot. Whether the actual whole-house cooling load + duct losses actually adds up to 48K is another question. If your cooling load is bigger than that you probably would have noticed- is it running all day and the house is still climbing into the mid 80s? Odds are you'll be able to back off to 3 tons or even 2.5 tons going ductless, but it's important to run the numbers. I have less of a feel for the true average cooling loads in TX for a house that size, so it's important to run the numbers using a real heat loss tool lest I steer you wrong, but 4 tons would seem a bit high for a house with ducts INSIDE of conditioned space, higher still for ductless. (I'm usually having to pay more to their heating output at lower temperaturs than cooling capacity.)
The
99th percentile outside design temperature for Victoria TX is 94F, but has a hefty 44 grains of water/lb latent load @ 50% RH. You might try running a room-by-room heating & cooling load calc using
Taco's freebie download. It's more of pro-type tool and delivers both heating and cooling loads, including latent loads, and tots up the whole-house loads. The default outside design temps in the tool for my location are substantially different from the ACCA Manual-J data (the first link) so change them to the ACCA values to keep from oversizing. If any of the bedrooms show up as substantially higher than the other separate rooms you might consider a 2 or 3 head multi-split, but barring substantial solar gain the temperature differences between rooms will usually be less than 5F at design temp, less than 3F at average mid-summer or mid-winter conditions if there's decent room for air to convect in hallways, etc.
If the heating load at the 28F design temp may or may not be higher than the 94F sensible cooling load or conversely it makes sense to upsize to a unit that covers 100% or even 125%+ of the load at design temp. Since these are modulating systems with a 3:1 turn-down ratio or more there's no efficiency or comfort penalty for oversizing up to about 150% oversizing in heating mode, at which point the wind-chill of the higher-cfm head for an oversized unit starts to take a toll. Unlike traditional heat pumps there's even an efficiency GAIN with modest oversizing, since their efficiency at part-load is higher than when running full-out (mostly due to the high efficiency of the ECM motors at any speed.) With oversized heat exchangers (for load) and slower compressor & blower drive, at 150% oversizing you gain about a full COP over what you'd get with a unit with output at design condition sized exactly for the load a that temp. So in heating mode instead of averaging ~4 in your location you'd be at 4.5+ if the unit is slightly oversized, with similar gains in cooling mode. This isn't something that necessarily shows up in an HSPF or SEER number (but units with higher numbers perform generally better than those with lower numbers.)