NVIDIA Ising is introduced as the first family of open-source quantum AI models designed to accelerate the development of quantum processors capable of running useful applications by tackling two key challenges: quantum processor calibration and quantum error correction. The model family, named after the mathematical Ising model used to simplify complex physical systems, provides high-performance and scalable AI tools that deliver “up to 2.5x faster performance and 3x higher accuracy” in decoding tasks for quantum error correction compared to traditional methods such as pyMatching. Ising positions AI as the control layer of hybrid quantum‑classical systems, enabling fragile qubits to be transformed into scalable and reliable quantum‑GPU platforms.
The Ising portfolio includes Ising Calibration, a vision‑language model that rapidly interprets measurements from quantum processors and automates continuous calibration, helping reduce calibration times from days to hours. It also offers Ising Decoding, composed of two 3D convolutional neural network variants optimized either for speed or for accuracy to perform real‑time quantum error‑correction decoding. These capabilities are framed within a broader market context in which quantum computing is projected to exceed 11 billion dollars by 2030, with progress depending on solving error‑correction and scalability issues.
A wide ecosystem of quantum enterprises, universities and laboratories, including Atom Computing, Academia Sinica, Fermi National Accelerator Laboratory, Harvard, Infleqtion, IQM, Lawrence Berkeley National Laboratory’s Advanced Quantum Testbed, Q‑CTRL and various U.S., European and Asian universities, has begun adopting Ising Calibration and Ising Decoding in research and development workflows. NVIDIA complements Ising with workflow “cookbooks,” training data and NIM microservices to simplify fine‑tuning for specific hardware, as well as integration with CUDA‑Q and NVQLink to support real‑time control and error correction in hybrid quantum‑classical architectures. Ising also joins a broader open‑model portfolio—such as Nemotron, Cosmos, Isaac and BioNeMo—available through platforms like GitHub, Hugging Face and build.nvidia.com, reinforcing NVIDIA’s strategy of open, high‑performance AI models for advanced computing domains.
Reference
NVIDIA. (2026, April 14). NVIDIA launches Ising, the world’s first open AI models to accelerate the path to useful quantum computers. NVIDIA Newsroom. https://nvidianews.nvidia.com/news/nvidia-launches-ising-the-worlds-first-open-ai-models-to-accelerate-the-path-to-useful-quantum-computers