Qiskit 2.1 Boosts Performance, Updates C API, Drops Python 3.9

IBM’s latest Qiskit v2.1 release prioritises performance and compatibility, rewriting core components of its Solovay-Kitaev decomposition in Rust to accelerate quantum circuit optimisation. The update introduces a new default QPY file version (15) and necessitates the optional ‘qpy-compat’ dependency for users reliant on files generated by older versions of the software. Developers utilising the C API must adapt to revised complex number handling, while those employing deprecated QuantumCircuit subclasses have until Qiskit v3.0 to migrate, ensuring continued functionality beyond the removal scheduled for approximately twelve months from now.

Performance and Efficiency Updates

Performance enhancements in Qiskit v2.1 centre on algorithmic optimisation and underlying code improvements. The Solovay-Kitaev decomposition, a crucial component in approximating arbitrary quantum gates with a universal gate set, has been reimplemented in Rust, yielding both faster execution speeds and improved precision. The default settings now utilise a depth of 12 and 5 repetitions.

Further optimisation is evident in the Sabre layout heuristics, employed during quantum circuit compilation to map logical qubits to physical qubit connectivity. Adjustments to the weighting of heuristic components within Sabre have resulted in increased routing efficiency.

Users maintaining extensive circuit archives should note alterations to the QPY file format. The default version emitted by qpy.dump() is now 15. While backwards compatibility is largely preserved, loading older QPY files requires installation of the optional qpy-compat dependency. A comprehensive Qiskit upgrade guide should incorporate these format considerations.

Finally, improvements extend to gate synthesis; the synthesis of 3-control MCX gates has been optimised to reduce the number of qubits required for their implementation, contributing to more efficient circuit execution.

New Feature Implementations

The C Application Programming Interface (API) has undergone revisions concerning the handling of complex numbers. The implementation now utilises a QkComplex64 struct, enhancing compatibility across a wider range of compilers and providing a more standardised approach. Developers utilising the C API should review these changes to ensure continued functionality.

Several bug fixes address specific issues within the Qiskit framework. Rendering problems affecting labels in the dag_drawer and DAGCircuit.draw() methods have been resolved. A bug preventing the correct ignoring of unused string-based parameters during assign_parameters() has been corrected. Users on Windows systems will find that edge cases in the Makefile configuration have been addressed. A minor correction has been made to the name attribute of the OrGate, which was previously incorrectly labelled as “and”. Issues related to deserialising circuits with reassigned parameters in qpy.load() have been fixed, and the synthesis of 3-control MCX gates has been optimised to reduce qubit requirements. Rendering of BoxOp instructions in the mpl mode of QuantumCircuit.draw() has also been improved.

API changes extend beyond the C API and QPY format. Layout plugins within present pass managers no longer contain their principal component if no coupling constraints are provided, affecting the behaviour of circuit optimisation passes and potentially requiring adjustments to custom pass configurations.

Users planning to upgrade to Qiskit v2.1 should be aware of several important considerations. Support for Python 3.9 will be removed in version 2.3, necessitating an upgrade to a later Python version. Reliance on older QPY files requires the installation of the qpy-compat extra dependency. Furthermore, numerous QuantumCircuit subclasses have been deprecated and will be removed in version 3.0, requiring migration to alternative implementations. A detailed Qiskit upgrade guide should incorporate these deprecation warnings to facilitate a smooth transition.

Deprecations and Compatibility Changes

Several API and architectural changes necessitate careful consideration during upgrades. Support for Python 3.9 is scheduled for removal in Qiskit v2.3; users continuing to utilise this version will require migration to a supported Python release. The deprecation of numerous QuantumCircuit subclasses, slated for removal in Qiskit v3.0, demands proactive migration to alternative circuit construction methods. A comprehensive Qiskit upgrade guide should highlight these specific classes to facilitate a planned transition.

Furthermore, the BlueprintCircuit base class is now deprecated, signalling its eventual removal from the framework. Users employing this class should investigate and adopt alternative approaches to circuit modelling. These deprecations, coupled with the changes to the C API and QPY format, underscore the importance of thorough testing following any upgrade to ensure continued functionality and compatibility.

More information
External Link: Click Here For More

Quantum News

Quantum News

As the Official Quantum Dog (or hound) by role is to dig out the latest nuggets of quantum goodness. There is so much happening right now in the field of technology, whether AI or the march of robots. But Quantum occupies a special space. Quite literally a special space. A Hilbert space infact, haha! Here I try to provide some of the news that might be considered breaking news in the Quantum Computing space.

Latest Posts by Quantum News:

IBM Remembers Lou Gerstner, CEO Who Reshaped Company in the 1990s

IBM Remembers Lou Gerstner, CEO Who Reshaped Company in the 1990s

December 29, 2025
Optical Tweezers Scale to 6,100 Qubits with 99.99% Imaging Survival

Optical Tweezers Scale to 6,100 Qubits with 99.99% Imaging Survival

December 28, 2025
Rosatom & Moscow State University Develop 72-Qubit Quantum Computer Prototype

Rosatom & Moscow State University Develop 72-Qubit Quantum Computer Prototype

December 27, 2025