Updated
Updated · OpenAI · Nov 3
OpenAI Expands AI Engineering Tools as Hardware Team Usage Jumps 23-Fold
Updated
Updated · OpenAI · Nov 3

OpenAI Expands AI Engineering Tools as Hardware Team Usage Jumps 23-Fold

3 articles · Updated · OpenAI · Nov 3

Summary

  • OpenAI said its latest models now support physical product design, simulation and manufacturing workflows across mechanical, electrical and system engineering, extending AI use beyond software into hardware development.
  • Two OpenAI staff used GPT-6 Astra to design a Codex-linked device in under a week, and a contractor produced a small prototype batch in under a month.
  • The test still exposed limits: Astra needed several review rounds and struggled with production constraints, tolerances and assembly details needed for manufacturing readiness.
  • On BenchCAD 2.0, a newer internal model scored 0.875 versus 0.594 for GPT-6 Astra, while token-based usage across OpenAI hardware teams has risen 23-fold since February.
  • OpenAI is widening industry trials with Chip Ganassi Racing and launching a Mazda collaboration on CAD, prototyping, aerodynamics and crash-safety work, while seeking more engineering partners.

Insights

Will OpenAI's aggressive expansion into robotics and crash safety testing finally bridge the gap between software and physical engineering?
How did an AI design a manufacturable physical prototype in just one week without any direct human editing?
Why do visually perfect AI-generated 3D components still fail complex assembly tests in real-world manufacturing?