Q.ANT announced its open-source toolkit on September 23. Its official repository provides C/C++ and Python interfaces and a CPU backend usable without Q.ANT hardware. Planned cloud hardware access is a separate future step.
GLOBAL IMAGING BRIEF
Analysis: accessible software lets developers explore workloads before committing to specialized hardware. A CPU development backend does not demonstrate a photonic processor's measured speed, precision or energy advantage.
Full perspective
The public repository separates the hardware-dependent package from the standalone CPU backend. That distinction matters when comparing what a developer can test today with the supplier's future deployment plans.
A useful evaluation should include data transfer, numerical error, supported operations, software integration and total system power, not only a favorable compute kernel. Developers can begin with reproducibility tests and baseline CPU or accelerator implementations. Any performance advantage should then be demonstrated on the intended hardware and workload with comparable precision and measurement boundaries.
Original source
Sources: Q.ANT's September 23, 2026 company-issued announcement and its official GitHub repository, checked September 28. No independent photonic-hardware benchmark is claimed.