GPT‑5.6 Sol Runs Quantum Chip Calibration at MIT Through Codex

Description and use

An AI agent (GPT‑5.6 Sol via Codex) automated routine quantum‑chip calibration tasks at MIT, selecting measurement parameters, operating lab hardware, and analyzing results, freeing researchers for higher‑level work.

Problem

Routine quantum‑chip calibration required repetitive manual measurements, parameter selection, and data analysis, consuming significant researcher time.

Outcome

The agent automated a substantial portion of a defined calibration workflow, freeing researchers to focus on experiment design and higher‑level analysis.

Workflow

• Identify qubit transition frequencies
• Calibrate control and readout pulses
• Estimate coherence times
• Run measurements on lab hardware
• Analyze measurement data
• Determine next calibration actions
• Escalate ambiguous data to human

Evidence

OpenAI says the Codex‑enabled agent handled an end‑to‑end set of clearly defined tasks, including identifying qubit transition frequencies, calibrating control and readout pulses, and estimating coherence.

Source

The agent selected measurement parameters, operated lab hardware, analysed results and chose subsequent steps for routine measurements on a superconducting six‑qubit chip.

Source