Three-SAT is SAT restricted so each clause has at most three literals:
The form (1) is still NP-complete (Karp), and in fact most modern reductions go through 3-SAT rather than general SAT — the fixed clause width simplifies gadget construction.
The 2-SAT version — same as (1) but with two literals per clause — is in P, solvable in linear time via the strongly-connected components of the implication graph. The boundary between 2 and 3 is one of complexity theory's sharpest cliffs:
(2) is the reason 3-SAT motivates entire industries: modern CDCL solvers, the Lasserre hierarchy, the Unique Games Conjecture.