OpenAI's GPT-5.6 Sol Automates Quantum Chip Calibration Workflow at MIT

OpenAI's GPT-5.6 Sol model, accessed via its Codex platform, was used to automate a significant portion of routine quantum chip calibration at MIT's Engineering Quantum Systems Group. Graduate student Beatriz Yankelevich deployed the AI agent to coordinate measurements on a six-qubit superconducting chip, handling tasks such as selecting measurement parameters, operating lab hardware, and determining next experimental steps. The agent managed an end-to-end calibration sequence — including identifying qubit frequencies and estimating coherence — that would otherwise require repeated manual effort from the researcher. OpenAI's case study clarifies that the system did not replace the physicist, but freed her to focus on higher-level tasks like experiment design and data analysis. The demonstration also acknowledges limits: noisy or ambiguous data may still require human judgment, and the agent automated a defined workflow rather than the broader scientific research process.