# Qiskit Fortran - Initial Release

**URL:** <https://fortran-lang.discourse.group/t/qiskit-fortran-initial-release/10937>\
**Category:** Announcements\
**Created:** [June 1, 2026, 1:43pm UTC](https://fortran-lang.discourse.group/t/qiskit-fortran-initial-release/10937 "2026-06-01T13:43:10Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![aaryav-3](https://avatars.discourse-cdn.com/v4/letter/a/e5b9ba/32.png) [@aaryav-3](https://fortran-lang.discourse.group/u/aaryav-3)\
**Post date:** [June 1, 2026, 1:43pm UTC](https://fortran-lang.discourse.group/t/qiskit-fortran-initial-release/10937/1 "2026-06-01T13:43:11Z")

</div>

Following [our earlier thread](https://fortran-lang.discourse.group/t/quantum-programming-using-fortran/10776) the initial release is live:

> **[GitHub - Qiskit/qiskit-fortran: Fortran binding for Qiskit](https://github.com/Qiskit/qiskit-fortran)**
>
> Fortran binding for Qiskit

The two-layer architecture sketched in that discussion became the skeleton of what’s here; a thin `ISO_C_BINDING` FFI layer over the [Qiskit C API](https://docs.quantum.ibm.com/api/qiskit-c) with an idiomatic Fortran module on top. Verified on macOS arm64 with gfortran 11–15 and flang 22.1.4; Linux is untested but expected to work.  
The point of the release is to put something real and functional and build from here together. With that in mind, we’ve tried to create entry points for everyone:

**For researchers and domain enthusiasts** , there’s an [`applications`](https://github.com/Qiskit/qiskit-fortran/tree/main/applications) directory for self-contained Fortran programs built on the binding. Building and running any application is a single configured [CLI instruction](https://github.com/Qiskit/qiskit-fortran/blob/a4dbcc6f9d332bd7f0b6b743a8fac035f171f31f/applications/README.md?plain=1#L12) against two paths for ease of use.  
A companion [`APPLICATIONS.md`](https://github.com/Qiskit/qiskit-fortran/blob/main/APPLICATIONS.md) outlines domains we think are natural fits from an initial search, not domain expertise (edits to this markdown more than welcome).  
If you work in any of these areas and see a demonstration worth building, the `applications` directory is exactly the right place for it, and the [application issues](https://github.com/Qiskit/qiskit-fortran/issues?q=label%3Aapplication) have some starter ideas already scoped to the current API surface.

One constraint to note at the current stage: `qiskit-fortran` has no execution runtime. Circuit construction and transpilation are fully native to the binding. Running a circuit against hardware or a simulator requires wrapping runtime code with this binding, as raised [here](https://github.com/Qiskit/qiskit-fortran/issues/12)

**For developer contributors** , there are [open issues](https://github.com/Qiskit/qiskit-fortran/issues) across the binding layer: fpm migration, `QkObs` (SparseObservable) bindings, transpiler module, CI, Linux validation, an open PR exploring [SWIG `%fortranbindc`](https://github.com/Qiskit/qiskit-fortran/pull/1) for FFI auto-generation, etc. Architectural questions have been formulated as open issues, to invite discussion.  
The design choices made so far are a starting point and are not settled. Thoughts on architecture, domain fit, or anything that looks wrong from a Fortran perspective are all welcome in the repo (as issues worth discussing/fixing).

Thanks especially to @hkvzjal, @gronki, @RonShepard, and @eelis ; the prior thread shaped a lot of what’s here.