IBM has issued a new release of Qiskit, version 2.2. The most significant feature of this new release is a continued increase in support for C that introduces a standalone transpiler function. IBM first introduced a C-language API with Qiskit 2.0 in April of this year that only provided interactions with a small number of Qiskit functions. But release 2.1 and this new release now greatly expands the C API and makes it possible for users to construct end-to-end quantum workflows that you can execute natively in C as well as other languages integrated via the C API.
This C compatibility is important in order to support IBM’s vision of Quantum Centric Supercomputing Centers (QCSC) for high performance computing (HPC). This is very important because a large part of HPC programming today uses the C++ or Fortran which is interoperable with this C API interface. So HPC programmers would no longer need to convert to a Python based language to use Qiskit. This also makes it easier to create hybrid Quantum plus classical HPC workflows.
In additions, there are several other new functions and improvements in this latest release. IBM is indicating that this new version improves transpilation performance by 10-20% on average. This is mostly due to the conversion of additional internal computations from Python to the more efficient Rust language. They have included some angle bounds support for hardware backends that can support fractional rotation gates. And in preparation for future systems that can support fault-tolerant computing, they have added an optional transpilation pass called the Litinski Transformation that can create a set of gate outputs that could be more suitable for fault tolerant architectures.
IBM has made some additional minor bug fixes in this release and is requiring that Qiskit be used with more recent versions of Rust and Python. In addition, some circuit library functions are being deprecated in version 2.2 and will be completely removed in version 3.0. To see more details about this new version of Qiskit you can visit blogs posted on the IBM Quantum website here and here. Also, the Qiskit 2.2 release notes are available here and code for a demo of Qiskit’s C compatibility that shows end-to-end compiled hybrid quantum/classical computation to implement a Sample-based Quantum Diagonalization (SQD) algorithm can be accessed here.
October 23, 2025
Leave A Comment