BlockBeats News, July 28th, NVIDIA released the open-source AI model NVIDIA Ising Calibration 1.5, designed to automatically analyze Quantum Processing Unit (QPU) diagnostic data, autonomously determine device calibration schemes, and achieve quantum computer calibration process automation.

NVIDIA stated that Ising Calibration 1.5 is a Visual Language Model (VLM) designed specifically for quantum computing calibration scenarios. It can understand quantum chip experimental data, complete "zero-shot" analysis in the absence of historical cases, and perform Contextual Learning with related experimental samples to assist quantum devices in continuously optimizing their operational status.

In the QCalEval Quantum Calibration Benchmark test, Ising Calibration 1.5 averaged about 10% ahead of similar-scale open-source models in zero-shot reasoning ability. When using related experimental cases for Contextual Learning, it achieved a performance improvement of approximately 86.5% compared to the previous generation model, surpassing multiple open-source models and approaching the level of top-tier closed-source large models.

This model has 31 billion parameters, supports NVIDIA Grace Blackwell and Vera Rubin data center GPUs, and also introduces the NVFP4 quantized version, which can be deployed on a single consumer-grade GPU or NVIDIA DGX Spark, significantly reducing the threshold for quantum laboratory usage.

NVIDIA stated that Ising Calibration 1.5 training data comes from various quantum bit architectures, including superconducting qubits, quantum dots, ions, neutral atoms, helium surface electrons, etc., providing calibration capabilities for different types of quantum computing devices.

The industry believes that automated calibration is one of the key bottlenecks for the scale development of quantum computing. NVIDIA's introduction of an AI-driven quantum calibration tool this time signifies that AI models are beginning to extend from the traditional computing domain to the quantum hardware control layer, potentially becoming a critical part of the future quantum computing industry infrastructure.