Nvidia Releases World's First Open-Source Quantum AI Model ISING to Address Core Quantum Computing Pain Points

Nvidia has launched ISING, the world's first open-source quantum AI model, designed to help researchers and enterprises build practical quantum processors and solve two core challenges: error correction and processor calibration.

Named after the mathematical Ising model, ISING offers high-performance, scalable AI tools. Its quantum error correction decoding is up to 2.5x faster and 3x more precise than traditional solutions. The calibration feature uses a vision-language model to automate processes, reducing calibration time from days to hours. The decoding function, based on 3D CNNs, outperforms the pyMatching open-source standard.

Currently, ISING is used by universities like Cornell and the University of Chicago. Nvidia also provides quantum computing workflow guides, training data, and NIM microservices for fine-tuning. It supports local operation to protect data and works with CUDA-Q and NVQLink to build a quantum supercomputing system. The model is open-source and available on platforms like GitHub.