UBTECH Thinker-VLA

Thinker-VLA 1.0 · Family: Thinker

Vision-language-action model from the UBTECH Thinker family: combines visual perception, language instructions and action generation. Applied to industrial manipulation tasks.

✓ Active✓ Public access⚖ Open sourceVision-Language-Action model📁 Thinker

🏢UBTECH RoboticsProducer

Deployment:📱 On-device☁ Cloud

Overview

UBTECH Thinker-VLA is a vision-language-action (VLA) model from the Thinker family by UBTECH Robotics. It is the third, execution tier of the Thinker stack: it combines visual perception, natural-language instructions and action generation, adapting the robot movements as objects and surroundings change.

Applications and performance

The model is used in industrial handling and loading tasks. According to UBTECH it improves inference efficiency in industrial scenarios by 176%. Architecture details, parameter counts and license have not been publicly disclosed.

Classification

Vision-Language-Action model

Family: Thinker

Applications

Robotic manipulation Robot policy training

Access & deployment

On-deviceCloud

Weights: Open source

Key parameters

📥 Input: image, text, robot sensors

Robotics

Robot manipulationRobot controlVisual groundingObject affordance understandingMotion planning

Technical specification

License

Open source (deklaracja UBTECH; brak potwierdzonego publicznego repozytorium)

Hardware requirements

Not publicly disclosed. A VLA execution model in the action layer of UBTECH humanoids (on-device and cloud inference).

Modalities

⬇ Input

imagetextrobot_sensors

⬆ Output

robot_actionsmanipulator_controlmotion_trajectories

Capabilities and applications

Native model capabilities

Vision-language-action grounding

The ability of a VLA model to ground visual perception and a language instruction into a concrete physical robot action. The model understands the scene and intent, then generates an executable action sequence, closing the loop from observation to motion.

Category: robotics

Action conditioning

Controlling model generation via action signals (camera, robot pose, commands, speech) rather than text prompts alone.

Category: multimodal

Planning

Forming and executing action plans for complex tasks.

Category: planning

Multimodal understanding

Category: multimodal

Robotics

Robot manipulationRobot controlVisual groundingObject affordance understandingMotion planning

Application domains