A groundbreaking experiment conducted by Hefei-based Origin Quantum has demonstrated the potential of quantum computing to revolutionize artificial intelligence. The company successfully fine-tuned a billion-parameter AI model using its domestically developed superconducting quantum computer, Origin Wukong, marking the first real-world application of quantum technology in large-model tasks.
The achievement, verified by scientists at the Hefei Comprehensive National Science Center, revealed that reducing model parameters by 76% led to an 8.4% improvement in training efficiency. This breakthrough suggests quantum computing could enable smaller, more efficient AI models — a critical advancement as traditional computing approaches face growing limitations.
Chen Zhaoyun, a researcher involved in the project, explained: 'This demonstrates quantum hardware's current capabilities in supporting practical AI applications.' The 72-qubit Origin Wukong system has already processed over 350,000 computing tasks since its launch, serving users in 139 countries through its quantum cloud platform.
Industry analysts suggest this development could impact specialized AI applications in fields like healthcare diagnostics and financial modeling. The ability to adapt general-purpose models like DeepSeek for niche applications using quantum systems opens new possibilities for global businesses seeking computational edge.
