The one-line difference
An agonist switches a receptor on. An antagonist blocks it. Orexin agonists increase orexin signalling to promote wakefulness. Orexin antagonists reduce orexin signalling to permit sleep.
Same receptor. Opposite direction. Opposite indication.
The antagonists came first
Once orexin was established as the wake-stabilising signal, the fastest commercial application was not to add it but to take it away. Blocking orexin should let a person with insomnia fall asleep, and it does. Several dual orexin receptor antagonists — suvorexant, lemborexant and daridorexant among them — reached the market as insomnia treatments years before any agonist reached late-stage trials.
They are the mechanistic proof, running backwards, of everything the agonist class rests on. Blocking orexin reliably produces sleep in people who cannot sleep. It follows that supplying it should produce wakefulness in people who cannot stay awake.
Why the agonists took much longer
Blocking a receptor is easier medicinal chemistry than activating one. An antagonist needs to occupy a site and do nothing; an agonist has to bind and then reproduce the specific conformational change a natural peptide produces. Finding small molecules that do that at the orexin receptors, survive being swallowed, cross into the brain and remain safe with daily use took the better part of two decades.
Along the way one candidate failed publicly on safety — Takeda's TAK-994, stopped in Phase 2 after liver signals — before its successor became the first approved drug in the class.
How to tell them apart
The naming convention helps, though it is not perfect. Insomnia antagonists tend to end in -orexant: suvorexant, lemborexant, daridorexant. The agonists in development end in -orexton: oveporexton, alixorexton, cleminorexton, danavorexton.
One letter separates them, so read carefully. If the drug is prescribed for insomnia it is an antagonist. If it is being developed for narcolepsy, hypersomnia or fatigue it is an agonist.