Overview
- Nvidia launched the Ising model family on Tuesday, publishing open weights, data and ready-to-run microservices on GitHub, Hugging Face and its own site.
- The company says the models cut calibration from days to hours and make error-correction decoding up to 2.5 times faster and three times more accurate than the pyMatching standard, claims that await independent tests.
- Ising plugs into Nvidia’s CUDA-Q software and NVQLink hardware to coordinate real-time control between GPUs and quantum processor units, the quantum chips that run qubits.
- Early users span labs and companies including Harvard, Fermilab, Lawrence Berkeley National Laboratory’s Advanced Quantum Testbed, IonQ, IQM, the U.K. National Physical Laboratory and others.
- Following Tuesday’s announcement, Nvidia shares rose and several quantum-computing stocks rallied, reflecting investor interest in AI tools that could stabilize today’s error-prone quantum systems.