On-Demand Quantum Runtime

Not every quantum computing project requires a dedicated QPU or a multi-block reservation. Researchers exploring new algorithmic ideas, engineers benchmarking quantum approaches against classical baselines, startups validating quantum-enhanced product concepts, and academic teams running coursework exercises all share a common need: fast, affordable access to production-quality quantum hardware without long-term commitments or complex procurement processes. The k&z On-Demand Quantum Runtime delivers exactly this — a fully managed, API-accessible quantum computing service billed by actual QPU consumption, with no minimum commitments, no reservation requirements, and no upfront fees.

The On-Demand Runtime operates on the same k&z QPU Block hardware that powers our dedicated and hybrid products. You are not accessing a lower-tier or prototype system; you are running circuits on production superconducting transmon qubits with the same gate fidelities, coherence times, and error-mitigation capabilities available to our largest enterprise clients. The difference is the access model: instead of reserving exclusive hardware for fixed time windows, you submit circuits to a managed queue that multiplexes workloads across a pool of QPU Blocks. The system is designed so that queue times remain under 30 seconds for most circuit submissions, even during peak demand periods, ensuring that your interactive development workflow is never blocked by long wait times.

Billing is transparent and granular. You pay for three dimensions of consumption: QPU wall-clock time (the duration your circuits occupy the processor, measured in seconds), shot count (the number of times each circuit is executed to build measurement statistics), and classical processing time (any server-side compilation, transpilation, or error mitigation applied to your jobs). Each dimension has a published per-unit rate, and your real-time consumption is visible on the k&z dashboard at all times. There are no hidden fees for data transfer, API calls, result storage, or account maintenance. Most exploratory workloads — running a few dozen circuits with 1,000-10,000 shots each — cost less than running equivalent simulations on rented GPU instances, making real quantum hardware accessible at a price point that was previously unimaginable.

Access is provided through the k&z Quantum API, a RESTful service that accepts circuit descriptions in OpenQASM 3.0, Qiskit circuit objects (serialized via QPY), Cirq JSON format, or our native QASM-Z intermediate representation. The API supports synchronous execution (submit and wait for results) and asynchronous execution (submit, receive a job ID, poll or webhook for completion). Client libraries are available for Python, JavaScript/TypeScript, Go, Rust, and C++, with full support for popular quantum SDKs including Qiskit, Cirq, Pennylane, and TKET. Authentication uses API keys with fine-grained IAM policies, enabling organizations to issue scoped credentials to individual researchers or CI/CD pipelines.

Every circuit submitted through the On-Demand Runtime passes through k&z's Meridian error-mitigation stack by default, applying zero-noise extrapolation and dynamical decoupling to improve result quality at no additional cost. For users who require more aggressive mitigation — such as probabilistic error cancellation, which increases shot overhead — optional premium mitigation tiers are available at published per-shot surcharges. Users can also disable Meridian entirely for raw hardware benchmarking or when implementing custom mitigation strategies.

Key Capabilities

Technical Specifications

Parameter	Specification
Available Qubit Counts	Up to 256 qubits per circuit (single QPU Block)
Maximum Circuit Depth	No hard limit; practical limit ~2,000 two-qubit gates with Meridian
Maximum Shots per Job	1,000,000 (configurable per circuit)
Median Queue Time	< 30 seconds (standard tier); < 5 seconds (priority tier)
Circuit Input Formats	OpenQASM 3.0, Qiskit QPY, Cirq JSON, QASM-Z, Pennylane
API Protocol	REST (HTTPS) with WebSocket streaming for real-time results
Client SDKs	Python, JavaScript/TypeScript, Go, Rust, C++
Authentication	API keys with IAM role-based access control
Billing Dimensions	QPU seconds + shot count + classical processing seconds
Minimum Charge	None (no minimums, no monthly fees)
Error Mitigation (Included)	Zero-noise extrapolation + dynamical decoupling (Meridian Base)
Error Mitigation (Premium)	Probabilistic error cancellation, measurement error mitigation
Mid-Circuit Measurement	Supported (feed-forward latency < 500 ns)
Result Retention	90 days (standard); extended retention available
Global Availability	API endpoints in North America, Europe, and Asia-Pacific

Ideal For

• Proof-of-concept development — Quickly validate whether a quantum approach offers advantage for your specific problem instance before committing to dedicated hardware. Run exploratory circuits, compare quantum results to classical baselines, and build the evidence base needed to justify larger investments in quantum infrastructure.
• Algorithm benchmarking and research — Evaluate quantum algorithm performance on real hardware across different circuit depths, qubit counts, and error-mitigation strategies. The consumption-based model lets you run extensive benchmark suites without worrying about wasted reservation time during analysis and development periods between experimental runs.
• Academic teaching and coursework — Provide students with hands-on access to production quantum hardware at costs that fit departmental budgets. The API-based access model integrates naturally with Jupyter notebooks and course management systems, enabling interactive quantum computing labs without specialized IT infrastructure.
• Startup and early-stage development — Quantum-focused startups building applications, middleware, or tools can integrate k&z hardware into their products via the API without capital expenditure on dedicated QPU access. Scale consumption as your customer base grows, paying only for the quantum resources your platform actually uses.
• CI/CD integration and automated testing — Incorporate quantum hardware execution into continuous integration pipelines using API keys scoped to CI service accounts. Automatically validate quantum circuit correctness, monitor hardware-aware performance regression, and generate benchmark reports with every code commit.
• Quantum software development — Test compilers, transpilers, optimizers, and error-mitigation techniques against real hardware behavior. The On-Demand Runtime provides the rapid iteration cycle that quantum software development requires — submit a modified circuit, observe the hardware response, adjust, and resubmit, all within minutes rather than hours.

The k&z On-Demand Quantum Runtime removes every barrier between your quantum idea and real hardware execution. There is no procurement paperwork, no capacity planning, no infrastructure management, and no financial risk. You write a circuit, submit it through the API, and receive measurement results from a production superconducting quantum processor — billed at a per-second rate that makes experimental quantum computing financially accessible to individual researchers, small teams, and organizations of any size.

Getting started takes minutes. Visit the k&z developer portal to create an account, generate your first API key, and run the quickstart tutorial that walks you through submitting a Bell state circuit and interpreting the measurement results. From there, explore our example notebooks covering quantum chemistry, optimization, machine learning, and error correction — all executable on real hardware through the On-Demand Runtime with a single API call.