OpenAI has moved robotics researchers into other groups because general-purpose models might render specialised robotics AI unnecessary.
The company is not abandoning the field. Instead, it is accelerating progress by integrating staff into the main research team led by Chief Research Officer Mark Chen. Researchers working on applied robotics and hardware now report to Greg Brockman’s Applied organization.
OpenAI made this decision on short notice. It had already booked a booth at the Conference on Robot Learning in Austin, which the venue reports sat empty. The company believes models like GPT-6 Astra can control robots directly, cutting out the need for specialised robotics software.
An Nvidia study showed an Astra-powered agent achieved a 74.8 percent success rate on simulated robotics tasks. The best specialised models managed 51 percent. Tests used a simulated DROID robot rather than a physical one.
A new benchmark called StationeryBench compared GPT-6 Astra with Ai2’s MolmoAct2 on five tasks involving desk objects on dual-arm YAM robotic arms. Across 200 trials, Astra fully completed 7 out of 100 tasks. MolmoAct2 did not complete a single one. Astra’s average progress score was 46, compared to 12 for MolmoAct2. In another test, Astra running on a humanoid robot navigated and cleaned up a kitchen.
It is unclear how much robotics data OpenAI and others incorporated during training. Other large general-purpose models from Anthropic and Google have also made recent progress on robotics tasks. Specialised robotics models may still hold advantages in cost and reliability, speed and energy consumption, even if they do not generalise as well.
Yoav Artzi, an AI researcher at Cornell University and Google DeepMind, sees Astra as a major leap in spatial understanding. On the still-unpublished REMAP benchmark, GPT-6 Astra reaches near-human accuracy, according to Artzi. He suspects OpenAI may have trained the model on large amounts of 3D training data such as Blender scenes.
OpenAI and robots, it’s complicated
OpenAI shut down its robotics division in 2020 because co-founder Wojciech Zaremba stated the company lacked the training data it needed and large language models were a faster path to AGI. The company rebuilt the team in 2024. CEO Sam Altman is reportedly targeting a first public robot demo for 2027, and OpenAI wants to build its own everyday robots in the long term. The company also faces heavy financial pressure to focus on its core AI business, where it has lost significant ground to Anthropic.
What it means
For people building or using robots, the shift suggests a move away from customised control software towards broader models. If a single model can handle various tasks without specific training, developers might spend less time writing specialised code. However, the current tests rely on simulations. Real-world robots still face different challenges regarding cost, reliability, and energy use that general models have not yet solved.
