MIT researchers use GPT-5. 6 Sol to automate quantum experiments

Researchers at MIT’s Engineering Quantum Systems Group (EQuS) have successfully utilized OpenAI’s GPT-5. 6 Sol model to automate complex quantum computing experiments. By connecting the AI to laboratory software via Codex, graduate student Beatriz Yankelevich deployed an agent capable of autonomously conducting measurements on a superconducting six-qubit chip.

The AI-driven system manages repetitive and labor-intensive calibration workflows, including identifying qubit transition frequencies, calibrating control and readout pulses, and estimating coherence times. Because the system can analyze results and dynamically adjust experimental parameters, it can execute end-to-end tasks with minimal human intervention.

This automation allows researchers to shift their focus from routine, manual measurements to higher-level scientific activities such as experiment design and data interpretation. The deployment demonstrates a practical application of AI agents in scientific research, where the model operates within bounded, hardware-connected workflows while humans retain oversight for complex decision-making.