Commercial quantum annealing and computing provider D-Wave Quantum Inc. (NYSE: QBTS) has announced algorithmic upgrades to its Stride hybrid solver alongside the rollout of a three-course developer curriculum titled Applied Quantum Optimization. Available via the Leap quantum cloud platform, the enhanced solver couples classical compute resources with D-Wave’s quantum annealing hardware to accelerate solution times and improve solution quality across combinatorial optimization problems.
| D-Wave Stride Hybrid Solver: Key Benchmark Comparisons | ||
|---|---|---|
| Optimization Problem Class | Benchmark Workload Scale | Demonstrated Runtime & Quality Gains |
| Traveling Salesperson (TSP) | Up to 10,000 target locations | Averaged within 0.02% of optimal result in 15 seconds; > 200× closer to optimal vs. prior Stride version at matched runtime. |
| Large-Scale TSP | Over 10,000 target locations | Averaged within 3.8% of optimal baseline in 120 seconds; > 500× closer to optimal vs. prior solver generation. |
| Assembly Line Balancing | Simple line balancing workflows | Achieved higher-quality solutions in 15 seconds than the previous Stride solver reached in 120 seconds. |
| Developer Curriculum | 3-course Applied Optimization track | Provides hands-on instruction for use-case formulation, mathematical modeling, and direct deployment via Leap. |
Combinatorial optimization represents one of the primary commercial applications for quantum annealing, spanning routing logistics, manufacturing scheduling, and supply chain allocation. Rather than treating optimization purely through unassisted QPU embedding, the Stride hybrid engine partitions problem sub-graphs across classical computing clusters and quantum annealers. The algorithmic revision targets the core solver architecture to shorten time-to-solution across large variable landscapes, demonstrating an eightfold speedup on assembly line benchmarks. To lower developer barriers, the accompanying learning path provides instruction on formulating Quadratic Unconstrained Binary Optimization (QUBO) problems and integrating hybrid solvers into production software stacks.
The release coincides with academic evaluations of D-Wave’s technology in data-constrained healthcare machine learning. A study led by researchers at Florida Atlantic University (FAU) and published in Discover Quantum Science, titled Inferring lung cancer under limited learning data through gate-based and annealing-based quantum techniques, evaluated quantum classifiers under intentionally restricted training conditions. Utilizing QUBO-based feature selection to filter 16 clinical attributes down to 10 key symptoms, the researchers trained quantum models on only 108 patient samples while evaluating predictive accuracy on the remaining 432 out-of-sample records.
Executed across five quantum software frameworks and both gate-based and annealing modalities, an annealing-based quantum support vector machine (QSVM) achieved the highest performance in the benchmark. Running on physical D-Wave hardware (Advantage_system6 on the Zephyr topology), the QSVM achieved 91.90% classification accuracy on hardware and 94.31% in simulation, outperforming gate-based implementations across Qiskit, Cirq, and PennyLane. Benchmarked against gate-based execution on an IBM quantum processor on a reduced 50-sample set, the annealing architecture completed hardware execution orders of magnitude faster due to lower circuit-shot overheads.
The updated solver and training modules are available immediately on the D-Wave Leap Cloud Platform here, with commercial benchmarking details accessible in the Official Corporate Announcement here and the clinical study detailed in FAU’s Research Release here.
November 4, 2026