OpenAI pushes its models into hardware engineering
OpenAI says GPT-6 Astra designed a working device without direct human edits, and it is signing up engineering partners including Mazda and Chip Ganassi Racing to test its models on physical-product design.
OpenAI described how its models are being used to design, simulate and build physical products across mechanical, electrical and system-level engineering. The company is also asking engineers, firms and software vendors to help shape the work.
The main demonstration is a small device for talking to Codex. OpenAI says GPT-6 Astra wrote the functional requirements, then designed the enclosure and internal structures, picked components, laid out the printed circuit board, and wrote the firmware and software integrations, all without any direct human edits. Two staff members built it with Astra in under a week as a side project, and a contract manufacturer delivered prototypes in under a month. OpenAI acknowledges that getting the design ready for manufacturing still took several rounds of review, because Astra did not consistently account for production limits, tolerances or assembly details.
Inside OpenAI, the hardware and infrastructure teams use Codex and ChatGPT for CAD models, manufacturing drawings, simulation setup and design review, covering both robotics and compute infrastructure. Token usage across those teams has grown 23-fold since February.
On BenchCAD 2.0, which tests 3D part and assembly reconstruction and circuit-board connection identification, an unnamed newer internal model scored 0.875 on average. GPT-6 Astra scored 0.594 at its highest reasoning setting. OpenAI did not publish external comparisons.
The company is expanding its work with Chip Ganassi Racing, turning 3D scans of parts into editable CAD models and 2D drawings. It is also starting a research collaboration with Mazda's aerodynamics and crash-safety engineers, and says results from that work will come "in the coming months."