Full specifications
6 of 14 fields stated| Field | Value | Source and caveats |
|---|---|---|
| Release | ||
| Released | 14 October 2026 | [A] "Metzingen, October 14, 2026" A |
| Input and output | ||
| Model type | not stated | Note NEURA's own term: "Its Perception-Language-Action (PLA) architecture connects what a robot sees, hears and feels with what it has already done". NEURA's page: "Neuron AI does what a VLA does, and adds touch, sound and memory for what a robot cannot see." A |
| Inputs | camera depth touch audio language robot state | Press release: "the model encodes 3D vision, tactile and force data, audio, robot state and language, aligns them in one shared space over time, and turns them into a single stream of action". Neuron AI page, 3D vision: "Geometry, colour and position in 3D." "Point clouds and depth: the scene as space, not a flat image." 3D vision = camera and depth, tactile and force = touch. A |
| Controls | arms and hands | Neuron AI page: "Five NEURA robot embodiments (4NE-1 Gen 3, 4NE-1 Gen 4, a wheeled 4NE-1, and two further bimanual platforms), with different grippers and hands." B |
| Degrees of freedom controlled | not stated | |
| Size and speed | ||
| Parameters | not stated | |
| Control rate | not stated | |
| Where it runs and what it is built on | ||
| On robot or cloud | on the robot | Neuron AI page: "Neuron AI is deployed locally, for real-time inference on the robot." B |
| Base model | not stated | |
| Openness | ||
| Access | not stated | Note Neuron AI page: "Talk to our team about Neuron AI on your robots, your use case and your data." B |
| Licence terms | not stated | |
| Licence | not stated | |
| The developer’s own claims | ||
| Learns new tasks | needs retraining | Press release, on the platform Neuraverse: "It also gives customers tools to fine-tune the model for their own tasks, turning process expertise into robot skills." A |
| Works across robots | yes | Developer claim: "Five NEURA robot embodiments (4NE-1 Gen 3, 4NE-1 Gen 4, a wheeled 4NE-1, and two further bimanual platforms), with different grippers and hands." The page also says "one model, deployed on every embodiment". B Yes only if the developer itself says the model runs on different robot models, for example robots from two manufacturers or two of one manufacturer’s models. Several units of the same robot do not count. |
How to read the figures
- 50HzA
- The letter in the grey square after a figure points to the source the figure came from. Pages with several sources carry several letters.
Notes
- NEURA introduced Neuron AI in a press release dated 14 October 2026.
- NEURA Robotics lists its address as Metzingen, Germany.
Versions and notes in the key figures (4)
Model type
Note NEURA's own term: "Its Perception-Language-Action (PLA) architecture connects what a robot sees, hears and feels with what it has already done". NEURA's page: "Neuron AI does what a VLA does, and adds touch, sound and memory for what a robot cannot see."
Controls
Neuron AI page: "Five NEURA robot embodiments (4NE-1 Gen 3, 4NE-1 Gen 4, a wheeled 4NE-1, and two further bimanual platforms), with different grippers and hands."
Inputs
Press release: "the model encodes 3D vision, tactile and force data, audio, robot state and language, aligns them in one shared space over time, and turns them into a single stream of action". Neuron AI page, 3D vision: "Geometry, colour and position in 3D." "Point clouds and depth: the scene as space, not a flat image." 3D vision = camera and depth, tactile and force = touch.
Access
Note Neuron AI page: "Talk to our team about Neuron AI on your robots, your use case and your data."
Sources
-
A
Primary source. neura-robotics.
com/ · Retrievedneura-robotics-introduces-neuron-ai/ -
B
neura-robotics.
com/ · Retrievedneuron-ai/
One line per source, not per figure: a source has one retrieval date. The letter beside a figure points to the line it came from.