GENE-26.5
About
About GENE-26.5
GENE-26.5 is Genesis AI's foundation model for dexterous robotic manipulation. Genesis evaluates the model on physical robots through a low-latency control stack and trains it with a data engine built around tactile-glove demonstrations and human video. The same model weights underpin laboratory demonstrations of cooking, pipetting, wire harnessing, smoothie preparation, cube manipulation and multi-object sorting. These physical demonstrations require task-specific robot data and a configured robotic platform.
The training approach starts with diverse human demonstrations, then aligns them with robot-execution data. Genesis World provides simulation for closed-loop policy evaluation and iteration. A separately trained piano policy tests the control stack, while the 20-active-degree-of-freedom Genesis Hand 1.0 was introduced as a further hardware step. Genesis develops GENE for its general-purpose robot programme, with public model access and commercial terms still unspecified.
Capabilities
Source-backed capabilities
- Operating environments: Manufacturer laboratory demonstrations
- Control modes: Autonomous
- Food preparation
Demonstrated - 2026-05-07 - Crack eggs, cut tomatoes and cook using coordinated two-hand manipulation and kitchen tools. - environment: Manufacturer laboratory demonstrations
Source - Laboratory liquid handling
Demonstrated - 2026-05-07 - Fit a pipette tip, transfer liquid into a tube, seal the tube and place it in a centrifuge. - environment: Manufacturer laboratory demonstrations
Source - Wire-harness handling
Demonstrated - 2026-05-07 - Bundle cables with both hands, place them on a stand and wrap them in tape. - environment: Manufacturer laboratory demonstrations
Source
Research
- GENE-26.5: advancing robotic manipulation to human level
2026-05-06 - Research Publication - Genesis AI unveiled GENE-26.5, a robotics foundation model for dexterous physical manipulation. The company said the system combines a human-scale dexterous robotic hand with a data engine for skill transfer, training and evaluation.
Source
Context
Market context
GENE-26.
- Target environment: Robotics labs, manipulation research environments and Genesis AI data-collection systems.
- Workflow context: Dexterous manipulation, multi-step physical tasks, robot-hand control and embodied-AI learning.
- Customer context: Robotics researchers, AI labs and customers exploring foundation-model control for manipulation-heavy robots.
- Deployment model: Company model release tied to Genesis AI's proprietary data engine, hand hardware and demonstrations.
- Commercial maturity: Early public foundation-model release with technical material and high-profile demonstrations.
- Adoption constraints: Adoption depends on hardware access, data quality, real-world task reliability and independent evaluation.
- Market position: Robot foundation-model lane focused on manipulation rather than vehicle autonomy or generic datasets.
- Adjacent products: Helix, sudo R1, FluxVLA Engine, Project GR00T
Comparable Systems
- Being-H0.8
BeingBeyond - Robotics Foundation Model
- Helix
Figure AI - Robotics Foundation Model
- FusionMax
Muks Robotics AI Pvt. Ltd. - Robotics Foundation Model
- GEN-1
Generalist - Software
- Lumo 2
Astribot - Robotics Foundation Model
- Rho-alpha
Microsoft - Robotics Foundation Model
Related Systems
- DYNA-1
Dyna Robotics - Robotics Foundation Model