Toronto, Canada – July 16, 2026 — PkTron, the advanced quantum computing framework developed by CETQAP (Centre of Excellence for Technology Quantum and AI Pakistan), has officially announced the release of Version 9.0.6, introducing a major collection of compiler, runtime, hardware-aware optimization, noise learning, and circuit serialization capabilities.

The latest release focuses on making quantum software development faster, more intelligent, and more hardware-aware while maintaining full compatibility with existing PkTron workflows.

Smarter Circuit Compilation with DAG Infrastructure

One of the headline additions is PkDag alongside TranspileStage, providing developers with a complete Directed Acyclic Graph (DAG) representation of quantum circuits.

The new implementation supports:

• Topological circuit ordering
• Predecessor and successor queries
• In-place node substitution
• Custom transpiler pass development

The interface mirrors Qiskit’s C-API QkDag while remaining a fully native Python implementation, allowing researchers to build sophisticated compiler optimizations without reconstructing transpilation pipelines from scratch.

Hardware-Aware Quantum Targeting

Version 9.0.6 introduces CouplingMap and an expanded Target architecture, enabling hardware-aware compilation.

New capabilities include:

• Device connectivity mapping
• Breadth-first search distance calculations
• Neighbor discovery
• Per-gate error rates
• Per-qubit T1 and T2 coherence times
• Readout error characterization
• Hardware targets generated directly using Target.from_coupling_map(...)

These additions allow PkTron to make more informed optimization decisions based on actual device characteristics.

Advanced VF2 Layout Optimization

A significant compiler enhancement arrives through VF2Layout and VF2PostLayout.

Using VF2-style subgraph isomorphism, PkTron can now determine more optimal qubit mappings before execution.

After routing completes, the new VF2PostLayout pass evaluates alternative assignments using hardware error information from the Target object and automatically selects mappings with lower expected execution error.

This results in improved circuit reliability on noisy quantum hardware.

Clifford+T Gate Synthesis Arrives

Version 9.0.6 also debuts the GridsynthDecomposer, bringing Clifford+T synthesis for single-qubit rotations.

The decomposer provides:

• Exact synthesis for Clifford+T multiple angles
• Examples include:
  • π/4 → T
  • π/2 → S
• Approximately 0.10 worst-case operator-norm error for arbitrary rotations

Although intentionally implemented as a documented simplified alternative to the complete Ross-Selinger number-theoretic algorithm, every synthesized sequence is independently verified against the ideal Rz rotation before being returned.

New Runtime Execution System

PkTron now includes RuntimeExecutor, a dedicated runtime submission engine.

Unlike the existing AsyncExecutor, which serves as a thread-pool task runner, RuntimeExecutor introduces true quantum job management through:

• .status()
• .result()

This allows developers to execute arbitrary quantum programs against supported backends using a workflow similar to production quantum runtime services.

Continuous Noise Learning

Another notable addition is PauliNoiseLearnerV2.

Instead of rebuilding noise models from scratch after every calibration cycle, the new implementation supports incremental learning through:

• .update()
• Exponential Moving Average (EMA) refinement
• Continuous adaptation toward newly observed calibration data

This enables long-running systems to preserve historical calibration knowledge while gradually incorporating fresh hardware information.

Massive Reduction in Circuit Storage Size

Perhaps one of the most eye-catching improvements comes from QPYCodec and FastQPYCodec.

PkTron now supports exact binary circuit serialization alongside a deduplication-optimized format that stores repeated gates and repeated sub-circuits only once.

According to the included benchmark_qpy results, repetitive workloads can achieve approximately:

🚀 Up to 45× smaller serialized circuit files

Importantly, developers should not expect faster execution, as PkTron states serialization speed remains roughly comparable. The improvement focuses entirely on reducing payload size for storage and transmission.

Designed for Next-Generation Quantum Development

Version 9.0.6 represents another step toward making PkTron a comprehensive quantum software ecosystem by combining advanced compiler technology, hardware-aware optimization, runtime execution, adaptive noise learning, and efficient binary serialization into a unified platform.

The release continues PkTron’s focus on delivering research-oriented capabilities while providing practical tools for education, simulation, and future quantum hardware integration.

With Version 9.0.6, developers gain access to a more intelligent compilation pipeline, richer hardware modeling, improved circuit fidelity, and significantly more efficient circuit storage, strengthening PkTron’s position as one of the most feature-rich quantum development frameworks emerging from Pakistan.