Quantum Annealing 2 QPUs available

Quantum Annealing Quantum Computing

Quantum annealers are special-purpose quantum devices that solve optimization problems by slowly evolving a quantum system from an initial superposition state to the ground state of a problem Hamiltonian encoding the optimization objective. D-Wave systems use superconducting flux qubits in a programmable Ising model. Unlike gate-based QPUs, annealers do not perform universal quantum computation.

Operating Temp
~15 mK (dilution refrigerator)
Gate Speed
20 µs per anneal (full problem solve)
Typical Fidelity
Not applicable — stochastic sampling approach
Scalability
High qubit counts achieved

Key Advantage

Very large qubit counts (5000+ physical qubits), fast sampling times (~microseconds per anneal), and a well-developed software ecosystem (D-Wave Ocean SDK) optimized for combinatorial optimization problems in logistics, finance, and scheduling.

Key Challenge

Limited connectivity between qubits requires problem embedding that can consume many physical qubits to represent a single logical variable. Not universal — cannot run arbitrary quantum algorithms. Quantum advantage over classical optimization solvers has not been conclusively demonstrated.

Quantum Annealing QPUs (2)

QPU Qubits Best Price Link
D-Wave Advantage2
D-Wave
 4,400 From $0.3000/task Details
D-Wave Advantage
D-Wave
 5,000 From $0.3000/task Details

Use Cases

Combinatorial optimization Logistics and supply chain Financial portfolio optimization Drug discovery and molecular simulation Machine learning (training boltzmann machines)

Frequently Asked Questions

What is the difference between quantum annealing and gate-based quantum computing?
Gate-based QPUs perform universal quantum computation using sequences of quantum logic gates, analogous to classical logic circuits. Quantum annealers instead encode an optimization problem in a physical Hamiltonian and use quantum tunneling to find the ground state (minimum energy configuration). Annealers are purpose-built for combinatorial optimization and cannot run general quantum algorithms.
How many qubits do D-Wave annealers have?
The D-Wave Advantage system has 5,000+ qubits with a Pegasus connectivity graph (15 couplers per qubit). The newer D-Wave Advantage2 has 4,400 qubits with a Zephyr topology (20 couplers per qubit, better connectivity). More qubits allow larger problems to be embedded directly.
What is problem embedding in quantum annealing?
Quantum annealers have limited physical qubit connectivity (15–20 neighbors per qubit). Real-world problems often require logical qubits to be connected to many others. Embedding maps logical problem variables to chains of physical qubits. This overhead means a 5,000-qubit annealer may solve problems with only a few hundred logical variables.
What is D-Wave Leap pricing?
D-Wave Leap charges by QPU access time: $2,000 per minute of actual QPU time. Since each anneal takes only 20 microseconds, a typical job costs fractions of a cent for the QPU time itself. However, there is overhead for job submission and readout. The free tier offers 1 minute/month for academic users.
Has quantum annealing demonstrated quantum advantage over classical computing?
This remains an open research question. While D-Wave systems consistently outperform simulated annealing on certain benchmarks, well-optimized classical algorithms (Branch & Bound, specialized heuristics) remain competitive for most practically-sized problems. Quantum annealing may offer advantages for specific problem structures at sufficient scale.

Compare With Other Technologies

Quantum Annealing vs Superconducting

20 µs per anneal (full problem solve) gates vs 10–700 ns per gate

Compare D-Wave Advantage2 vs IBM Heron r2 →
Quantum Annealing vs Trapped Ion

20 µs per anneal (full problem solve) gates vs 1 µs – 1 ms per gate

Compare D-Wave Advantage2 vs Quantinuum H2-1 →
Quantum Annealing vs Neutral Atom

20 µs per anneal (full problem solve) gates vs 0.1 µs – 1 ms per gate

Compare D-Wave Advantage2 vs QuEra Aquila →
Quantum Annealing vs Photonic

20 µs per anneal (full problem solve) gates vs Picoseconds for passive operations; detector timing ~ns

Compare D-Wave Advantage2 vs Xanadu Borealis →