The issue with using a thermoelectric nuclear generator like the one on Curiosity is that there's nowhere to dump the heat. RTGs are all based on thermocouples to generate their power which requires a temperature delta between the RTG interior and exterior which is hard given the extreme temperatures of Venus. Maybe one based on thermovoltaic cells would work but those haven't been tried with RTGs yet.

Wind might be the best bet. Turn one of the dangers into a power source at least, designing one that would work for a long time in those conditions wouldn't be easy but maybe.

Forget RTGs; what about actual nuclear reactors? Obviously, this means a much larger probe. But even with Venus's high atmospheric temperature, a nuclear reactor's output is much higher, so theoretically it should be possible for it to work. Then, with a very large amount of power available, it should be possible to use active cooling for the electronics, at least until the nuclear fuel is depleted.

Can a nuclear reactor actually work if the steam never recondenses? Either way that's a long ways away, the smallest currently planned nuclear reactor is still the size of a shipping container. Building a probe around that would make and it's transfer stage the largest thing ever put in orbit by a long ways then there's the logistic of actually landing the thing.

IANANE (I am not a nuclear engineer), but I don't think you can use water here, for the reason you state. Water makes sense on Earth because the temperature and pressure range suits it, but on Venus I think you'd need to find some other fluid to transfer heat to the outside, perhaps liquid sodium? But nuclear reactors work the same way as any other heat engine: you have a source of heat and exploit a temperature differential by moving heat from the heat-source to the environment. So it should work just fine on Venus, but with some other working fluid.

You're right though: this would be rather large. However, it might be possible to make it smaller, by eliminating shielding. Your smallest planned reactor surely uses a bunch of shielding, because it's intended for use on Earth, around humans. For an automated probe on Venus, that's not a concern so you don't need shielding, except whatever's necessary for safe handling before launch. This is also a place where it'd be a lot better if we had a Moon base where we could build things like this; it's a lot easier to launch mass from the Moon than from the Earth. Landing the probe should be easy: Venus's atmosphere is extremely dense so a parachute (resistant to the sulfuric acid, at least for a short time) should be fine, though the high windspeeds could be a problem.