The paper introduces Qiskit, a software development kit for quantum information science, highlighting its design philosophy, architecture, and core components. Qiskit is designed to facilitate research, education, and practical applications in quantum computing. Key features include modularity and extensibility, a balance between performance and rapid prototyping, portability and hardware optimization, and interoperability across different abstraction levels. The software supports quantum-classical integration, allowing for real-time and near-time classical computations to enhance quantum circuit performance and error mitigation. The paper discusses the transpiler, which optimizes and translates quantum circuits to target instruction sets, and the pass manager, which orchestrates the transformation pipeline. Primitives, such as samplers and estimators, are used to evaluate circuits and perform common quantum computational tasks. An end-to-end workflow for solving a Hamiltonian simulation problem in condensed matter physics is demonstrated, showcasing Qiskit's capabilities in circuit representation, optimization, and execution. Qiskit has a thriving ecosystem of tools and plugins, and its flexibility has enabled its use in various quantum computing platforms and applications. The paper concludes by discussing the future directions of Qiskit, emphasizing its role in advancing quantum computing through software co-design with hardware.The paper introduces Qiskit, a software development kit for quantum information science, highlighting its design philosophy, architecture, and core components. Qiskit is designed to facilitate research, education, and practical applications in quantum computing. Key features include modularity and extensibility, a balance between performance and rapid prototyping, portability and hardware optimization, and interoperability across different abstraction levels. The software supports quantum-classical integration, allowing for real-time and near-time classical computations to enhance quantum circuit performance and error mitigation. The paper discusses the transpiler, which optimizes and translates quantum circuits to target instruction sets, and the pass manager, which orchestrates the transformation pipeline. Primitives, such as samplers and estimators, are used to evaluate circuits and perform common quantum computational tasks. An end-to-end workflow for solving a Hamiltonian simulation problem in condensed matter physics is demonstrated, showcasing Qiskit's capabilities in circuit representation, optimization, and execution. Qiskit has a thriving ecosystem of tools and plugins, and its flexibility has enabled its use in various quantum computing platforms and applications. The paper concludes by discussing the future directions of Qiskit, emphasizing its role in advancing quantum computing through software co-design with hardware.