A review of exactly how Genie Sim functions.|Resource: AGIBOT
AGIBOT Development Modern technology Co. today presented Genie Sim 3.0, a next-generation robotic simulation system powered by NVIDIA Isaac Sim, at CES 2026. The firm stated it provides a linked, open simulation operations that unites electronic property generation, scene generalization, information collection, automated assessment, and physics-based simulation in a solitary toolchain.
AGIBOT stated it made Genie Sim’s core component, Genie Sim Criteria, as a standard assessment system to develop a precise and reliable standards for personified knowledge.
Genie Sim 3.0 attracts from greater than 10,000 hours of artificial datasets, consisting of real-world robotic procedure circumstances. The system incorporates 3D repair with aesthetic generation to produce a high-fidelity simulation atmosphere. It additionally utilizes huge language version (LLM) modern technology to produce scenes and assessment metrics.
The assessment system covers greater than 200 jobs throughout over 100,000 circumstances to construct a detailed capacity account for designs. AGIBOT stated the system can speed up version growth, minimize dependence on physical equipment, and encourage development in personified knowledge.
The Shanghai, China-based firm was developed in February 2023. AGIBOT uses a series of humanoid robotics and mobile robotics for cleansing and solution setups.
In December, AGIBOT launched its upgraded humanoid, AGIBOT A2, in December. Later on that month, the firm turned out its 5,000 th mass-produced humanoid robotic at its manufacturing facility.
Genie Sim 3.0 offers a full-stack pipe for training and assessment
Exactly how Genie Sim’s LLM-driven scene generation functions.|Resource: AGIBOT
AGIBOT stated its incorporated innovations allow it to provide real-time, high-fidelity, and high-precision simulation atmospheres.
The system catches real-world atmospheres with Skyland Development’s MetaCam portable 3D laser scanner. This scanner incorporates high-resolution RGB photos, 360 ° lidar factor clouds, and centimeter-level RTK (real-time kinematic) placing. Interactable things can be exchanged simulation-ready properties from a solitary 60-second orbital video clip, dramatically speeding up scene building and construction, stated the firm.
A specifying attribute of Genie Sim 3.0 is its LLM-driven natural-language scene generation, AGIBOT declared. Individuals can define atmospheres conversationally, and the system instantly generates organized scenes, aesthetic sneak peeks, and countless semantic variants without hands-on reasoning coding. Vision-language designs even more fine-tune and tune these scenes to fulfill specification-level demands, allowing quick adjustment and solid version generalization, it declared.
Additionally, Genie Sim 3.0 incorporates genuine industrial-scene datasets right into both training and assessment pipes. This technique dramatically reduces algorithm-validation cycles and minimizes dependancy on equipment framework.
AGIBOT offers over 10,000 hours of artificial information
A review of exactly how Genie Sim takes care of multi-dimensional assessment.|Resource: AGIBOT
AGIBOT has actually additionally launched over 10,000 hours of open-source artificial information covering greater than 200 jobs with multi-sensor methods such as RGB-D, stereo vision, and whole-body kinematics. A smart data-collection toolkit sustains both low-latency teleoperation and automated job shows. With auto-annotation and a healing device that returns to collection after job failings, it can minimize the price and time of dataset manufacturing.
For assessment, Genie Sim 3.0 offers greater than 100,000 simulation circumstances, which AGIBOT stated relocations past single-metric standards to create full-spectrum capacity accounts for personified designs. Making use of LLMs, the system can auto-generate executable assessment process throughout varied semantic, spatial-reasoning, and adjustment measurements. The firm stated it details version toughness, constraints, and optimization paths.
All simulation properties, datasets, and automated assessment resource code are completely open-source. Designers and scientists can access AGIBOT Genie Sim on GitHub.

The message AGIBOT launches Genie Sim 3.0 robotic simulation system showed up initially on The Robotic Record.
发布者:Robot Talk,转转请注明出处:https://robotalks.cn/agibot-launches-genie-sim-3-0-robot-simulation-platform/