Quantum computing has become increasingly visible in pharmaceutical research, yet the evidentiary meaning of an “advantage” claim remains unsettled. Drug discovery contains heterogeneous scientific decisions, data structures, computational abstractions, validation standards, and translational consequences. Consequently, improvement in circuit execution, molecular-energy estimation, sampling, classification, or candidate generation cannot by itself establish pharmaceutical usefulness. This Original Quantum–Classical Position Article examines how quantum methods should be matched to problems they can credibly solve rather than evaluated through a single performance measure. It develops a proposed Quantum–Classical Credibility Map that links pharmaceutical problem specification, task selection, algorithmic assumptions, hardware conditions, data and encoding requirements, hybrid workflow design, comparator quality, uncertainty, and downstream validation. The synthesis distinguishes computational feasibility from practical advantage, scientific utility, and pharmaceutical value. It further separates electronic-structure calculations, reaction-mechanism analysis, molecular generation, property prediction, and quantum-assisted optimization according to their outputs, resource dependencies, and validation routes. The central proposition is that credibility is conjunctive: a claim cannot be stronger than the least-supported necessary connection between the pharmaceutical decision, computational subproblem, quantum method, classical workflow, and empirical evidence. The proposed framework does not establish quantum advantage, clinical utility, regulatory acceptance, or deployment readiness. Its value lies in organizing research questions, exposing hidden assumptions, improving quantum–classical comparisons, and defining conditions under which a pharmaceutical application should be investigated, reformulated, monitored, or deferred. Prospective benchmarking, independent replication, complete resource accounting, and experimentally connected evaluation are required before the framework itself or the applications organized through it can be considered validated.