Agibot · robot AI model
BFM-2
China · 2026
Sources retrieved: 2 November 2026 (A, B)
- What it controls
- Whole bodySource and note: What it controlsA agibot.com.cn/article/315/detail/161 · Retrieved 2 November 2026Note. AGIBOT's release article calls BFM-2 a whole-body motion foundation model that generates the full motion from the current state to the target state (translated from Chinese).
- Inputs
- Robot stateSource and note: InputsA agibot.com.cn/article/315/detail/161 · Retrieved 2 November 2026Note. AGIBOT's release article says BFM-2 combines the robot's whole-body dynamics state, its contact-constraint state and a high-level target command, and that a later BFM-3 is to add vision, touch, speech and other inputs (translated from Chinese).
Specifications
3/14 specifications disclosed 11 not stated
9 fields hidden 2 not-stated fields with a note stay ·
How to read the figuresShow guideClose
Not statedThe developer publishes nothing on it
Select a source marker for source details and notes.
Release1/1 disclosed
Released23 May 2026
Source and note: ReleasedB youtube.com/watch?v=jgo3TjxszF4 · Retrieved 2 November 2026Note. [A] “2026-05-23T00:19:26-07:00” Upload time of the video on AGIBOT's YouTube channel. AGIBOT's Chinese website dates its release article “2026-05-26 16:54:45”.
Input and output2/4 disclosed
Model typeNot stated
Source and note: Model typeB youtube.com/watch?v=jgo3TjxszF4 · Retrieved 2 November 2026Note. AGIBOT's own term is “a two-stage Motion-Between locomotion foundation model”; it does not call it a vision-language-action or world action model.
InputsRobot state
Source and note: InputsA agibot.com.cn/article/315/detail/161 · Retrieved 2 November 2026Note. AGIBOT's release article says BFM-2 combines the robot's whole-body dynamics state, its contact-constraint state and a high-level target command, and that a later BFM-3 is to add vision, touch, speech and other inputs (translated from Chinese).
What it controlsWhole body
Source and note: What it controlsA agibot.com.cn/article/315/detail/161 · Retrieved 2 November 2026Note. AGIBOT's release article calls BFM-2 a whole-body motion foundation model that generates the full motion from the current state to the target state (translated from Chinese).
Degrees of freedom controlledNot stated
Size and speed0/2 disclosed
ParametersNot stated
Control rateNot stated
Where it runs and what it is built on0/2 disclosed
On robot or cloudNot stated
Base modelNot stated
Openness0/3 disclosed
AccessNot stated
Licence documentNot stated
Licence typeNot stated
The developer’s own claims0/2 disclosed
Learns new tasksNot stated
Works across robotsNot stated
Source and note: Works across robotsYes 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.
Notes
AGIBOT describes BFM-2 as a two-stage Motion-Between locomotion foundation model.
AGIBOT says BFM-2 learns the general dynamics of whole-body states and is not limited to getting up after a fall.
Sources
One line per source, not per figure: a source has one retrieval date.
- Aagibot.com.cn/article/315/detail/161Main source. Retrieved 2 November 2026
- Byoutube.com/watch?v=jgo3TjxszF4Retrieved 2 November 2026