GPT-5.6 Sol, paired with OpenAI Codex, is now autonomously running quantum computing experiments at MIT. The system calibrates qubits, executes test sequences, and analyzes results without continuous human intervention.
The MIT researcher's workflow is the real story here. Codex handles the programming layer while Sol interprets experimental output and adjusts calibration parameters in response. That feedback loop, model driving experiment driving model, is what separates this from a demo.
The full piece details the specific calibration methods, the error rates Sol is working against, and where human oversight still enters the process. If you work anywhere near quantum hardware or autonomous research pipelines, the methodology section is worth your time.
[READ ORIGINAL →]