OpenAI reassigns robotics researchers to other teams
Original title: OpenAI reassigns robotics researchers to other teams as general-purpose models might make specialized robotics AI unnecessary
Staff move under Mark Chen and Greg Brockman after GPT-6 Astra beat specialized models on simulated robot tasks.
According to Fortune, OpenAI is dispersing its robotics group into existing organizations instead of ending robotics work, as it did in 2020. The move was so abrupt that a booked Conference on Robot Learning booth in Austin sat empty. The company is said to think general-purpose models can drive robots directly. A public robot demonstration is reportedly targeted for 2027, while OpenAI also faces pressure to focus on its core AI business.
Unconfirmed: The reassignment, 2027 demo target, and training-data claims are reported or suspected, not confirmed by OpenAI.
- An Nvidia study put Astra at 74.8 percent success versus 51 percent for the best specialized models.
- Those tests used a simulated DROID robot, not a physical robot.
- On StationeryBench, Astra finished 7 of 100 tasks; MolmoAct2 finished none.
- OpenAI shut its robotics division in 2020 and rebuilt the team in 2024.
- Artzi says unpublished REMAP results put Astra near human spatial accuracy.
- AI models
- GPT-6 AstraMolmoAct2
From the original article
2 min read at the-decoder.com
OpenAI is dissolving its standalone robotics team for the second time, but not because it lacks data. General-purpose models like GPT-6 Astra can apparently control robots directly, making a dedicated unit unnecessary. The researchers are being redistributed across other teams in what the company frames as a way to accelerate its robotics efforts.
By Matthias Bastian · the-decoder.com
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.