•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
NVIDIA has introduced two open-source AI models for quantum computing—Ising Calibration and Ising Decoding—designed to function as a next-generation operating system for quantum processors. The models focus on real-time calibration and error correction, aiming to improve how quantum processing units (QPUs) are controlled and stabilized during operation.
Ising Calibration is a 35-billion-parameter vision-language model trained to read experimental data from QPUs and propose the adjustments needed to improve performance. NVIDIA says that, when paired with autonomous agents, calibration time can be reduced from days to hours.
Ising Decoding is offered in two variants built on 3D convolutional neural networks, with 0.9 million and 1.8 million parameters. The models are optimized for speed and accuracy, respectively, and perform preliminary decoding within the surface-code quantum error correction mechanism.
In NVIDIA’s internal tests, Ising Decoding was reported to be approximately 2.5x faster and about 3x more accurate than the open-source pyMatching tool. NVIDIA also states that training data requirements were reduced by up to tenfold.
NVIDIA’s Sam Stanwyck, Director of Quantum Products, said that current quantum processors typically accumulate errors after about 1,000 operations. In that context, decoding speed can directly affect how long qubits remain stable. NVIDIA notes that a 2.5x improvement in decoding speed could extend the number of operations before significant errors occur.
Evaluations cited by Tom’s Hardware describe Ising as a notable step in NVIDIA’s expansion into quantum computing. The company does not manufacture QPUs, and instead emphasizes building the AI control and optimization layer intended to help move quantum technology toward practical applications.
Times of India contrasts Ising with chatbots such as ChatGPT or Claude, noting that Ising is designed to act on the data it receives rather than simply predicting the next word. The report also says Ising is built to operate in real time to avoid stressing the quantum computer, supporting faster processing and higher throughput.
Premium gym chains are entering a “golden era” that is ending or already in decline, as rising operating costs collide with shifting consumer preferences toward more flexible, community-based ways to exercise. Long-term memberships are shrinking, margins are pressured by higher rents and facility expenses, and competition from smaller, more personalized…
