A preview version with an explicit capability map
N/M's language version is 0.2.0; the coordinated Core, CLI, MCP and Language Server package release is 0.2.1. The overall product is a preview. That label should not be read as a promise that every feature has the same maturity: the parser and statevector runtime have stable entries in the capability manifest, while other language and toolchain features retain preview or experimental status.
Use the capability overview to check the current status before adopting a feature. New sessions enable a default language profile that includes several newer features; their default availability does not automatically change their maturity classification. Saved workspaces preserve explicit feature choices.
What you can explore today
The project includes a source editor, local execution, result inspection and development tools. A small circuit is a useful starting point:
module preview_rotation;
@seed(31);
fn main() {
let q = qreg[1];
sample 256 {
Ry(q[0], PI / 2);
let c = measure(q[0]);
}
return q;
}In the ideal statevector model, this rotation produces equal computational-basis probabilities. Run it in the Playground, then change the angle to 0 and to PI. Predict the resulting distributions before looking at the output.
Beyond basic circuits, the codebase provides numerical operations and bounded classical functions, learning examples, parameterized workflows and training tools. The workspace connects editing with diagnostics, simulation and artifact handling. The VS Code extension supplies language tooling outside the browser. Each of these surfaces has its own supported scope; a working simulator example does not establish support for every export target or every editor operation.
Choose a simulator by the problem
Statevector simulation stores amplitudes for the full quantum state and its memory demand grows exponentially with the number of qubits. Stabilizer simulation uses a different representation for supported Clifford circuits. Matrix-product-state methods have their own representation and bond-dimension constraints.
These are distinct tradeoffs, not interchangeable promises of unlimited qubit counts. Check the selected backend's supported operations, resource limits and approximation settings. A small example that succeeds locally is useful evidence for that example; it cannot establish the feasibility of a much larger circuit.
Keep versions and artifacts separate
The language version, package version and serialized artifact formats are different contracts. The 0.2.0 language does not require every intermediate representation, sidecar or workspace format to be numbered 0.2.0. Preserve the version information written into exported records, and use the documented importer rather than editing version fields by hand.
Core, CLI, MCP and Language Server 0.2.1 are available from public npm. The next tag selects the preview channel; use @0.2.1 to pin this release. An untagged install selects the separate latest channel. See the public installation guide and the CLI reference. The VS Code extension has its own distribution version and release history; consult the release page and extension guide. Free use does not grant an open-source license to modify or redistribute the package code; each installed package includes its distribution license.
A practical adoption checklist
- Start with a small reproducible circuit and inspect diagnostics before execution.
- Record the language version, source, target, seed and relevant feature settings.
- Check capability status for the operation you intend to use.
- Test export and re-import if portability is part of your workflow.
- Increase circuit size gradually while observing backend limits.
N/M is a quantum-software learning and experimentation toolchain. Local execution and an exported file are not evidence of execution on a physical QPU or a general quantum speedup.
Sources
- N/M language and capability overview: the product's capability statuses and tools.
- N/M documentation: supported workflows and language reference.
- Release records: component versions and release distinctions.
- IBM Quantum Learning: background on quantum states and circuits.