# Robot Control

Vision-language-action policies process hundreds of visual tokens per step, which drives up inference latency. APS-VLA keeps the backbone frozen and uses a lightweight scorer conditioned on the current observation, instruction and proprioception to pick tokens, training the retained representation both to reproduce policy outputs and to predict the subsequent visual change. On LIBERO with 128 tokens retained, mean success is 95.75%, 1.25 points below the unpruned 97.00%, and inference latency falls to 68.75 ms.

Vision-language-action policies process hundreds of visual tokens per step, which drives up inference latency. APS-VLA keeps the backbone frozen and uses a lightweight scorer conditioned on the current observation, instruction and proprioception to pick tokens, training the retained representation both to reproduce policy outputs and to predict the subsequent visual change. On LIBERO with 128 tokens retained, mean success is 95.75%, 1.25 points below the unpruned 97.00%, and inference latency falls to 68.75 ms.

Shanghai Jiao Tong University and collaborators assembled a roughly 47,000-hour multisource pool and extended a pretrained video model's language–video priors into language–video–action learning, yielding the world action model SWAM. On the 50 manipulation tasks of RoboTwin 2.0, SWAM reaches 61.56% success under the randomized Hard setting, 12.86 percentage points above the strongest baseline OpenWAM, and 88.10% under the clean Easy setting, slightly below OpenWAM's 89.30%. On a real bimanual platform, its three-task average Hard success rate is 40.00% against OpenWAM's 23.33%.

Shanghai Jiao Tong University and collaborators assembled a roughly 47,000-hour multisource pool and extended a pretrained video model's language–video priors into language–video–action learning, yielding the world action model SWAM. On the 50 manipulation tasks of RoboTwin 2.0, SWAM reaches 61.56% success under the randomized Hard setting, 12.86 percentage points above the strongest baseline OpenWAM, and 88.10% under the clean Easy setting, slightly below OpenWAM's 89.30%. On a real bimanual platform, its three-task average Hard success rate is 40.00% against OpenWAM's 23.33%.

Frontier models' spatial understanding now lets coding agents control a real robot arm zero-shot and end-to-end, with no task-specific demonstrations or retraining, so risk crosses from the digital world into the physical one. ArmedAgent sorts these risks by source into benign instructions that turn unsafe, malicious use, and indirect prompt injection, across 36 tasks in 5 scenarios, evaluated in simulation and validated on a physical arm. Testing GPT-6-Astra, GPT-6-Sol and GPT-6.1-Terra under Codex, overall attack success ranges from 72.22% to 94.29%, and refusal rate is 0% for every model in every category.

Frontier models' spatial understanding now lets coding agents control a real robot arm zero-shot and end-to-end, with no task-specific demonstrations or retraining, so risk crosses from the digital world into the physical one. ArmedAgent sorts these risks by source into benign instructions that turn unsafe, malicious use, and indirect prompt injection, across 36 tasks in 5 scenarios, evaluated in simulation and validated on a physical arm. Testing GPT-6-Astra, GPT-6-Sol and GPT-6.1-Terra under Codex, overall attack success ranges from 72.22% to 94.29%, and refusal rate is 0% for every model in every category.

Google DeepMind's Gemini Robotics ER 2, a high-level reasoning model for robots, is now in public preview on the Gemini platform, letting enterprises declare low-level controls such as vision-language-action models and navigation APIs as tools for the model to orchestrate. The model classifies task progress from continuous video and pinpoints critical event frames, with a stated 91.3% accuracy and 0.96-second mean absolute frame distance. It is described as working across embodiments from stationary arms to humanoids, with simulation, evaluation and deployment handled on Google Cloud.

Google DeepMind's Gemini Robotics ER 2, a high-level reasoning model for robots, is now in public preview on the Gemini platform, letting enterprises declare low-level controls such as vision-language-action models and navigation APIs as tools for the model to orchestrate. The model classifies task progress from continuous video and pinpoints critical event frames, with a stated 91.3% accuracy and 0.96-second mean absolute frame distance. It is described as working across embodiments from stationary arms to humanoids, with simulation, evaluation and deployment handled on Google Cloud.

Gemini 4 Argon puts physical understanding, code generation and visual-semantic reasoning into one model so robot tasks can be practiced in simulation first. On the Embodied Reasoning (ERQA) benchmark, which tests physical and spatial problems in 3D environments, it scores 87% using pure reasoning, above other frontier models. It can also reconstruct calibrated simulation twins from robot camera feeds, egocentric human video or hand-drawn sketches, and generate closed-loop control and training data from them.

Gemini 4 Argon puts physical understanding, code generation and visual-semantic reasoning into one model so robot tasks can be practiced in simulation first. On the Embodied Reasoning (ERQA) benchmark, which tests physical and spatial problems in 3D environments, it scores 87% using pure reasoning, above other frontier models. It can also reconstruct calibrated simulation twins from robot camera feeds, egocentric human video or hand-drawn sketches, and generate closed-loop control and training data from them.