Chinese robotics company AGIBOT used its Innovation Day 2026 in Shanghai on October 29 to ask a question that matters more as robots leave laboratories and enter factories, warehouses and service settings: how much productive work are robots actually delivering once they are deployed?
The company’s answer is Worken, a proposed framework for measuring the productive work performed by robots in real-world environments. The name combines “Work” and “Token”, and the idea is to shift attention away from how many robots are manufactured or delivered and toward how much useful work they can perform autonomously and consistently.
Key Takeaways
- AGIBOT introduced Worken at Innovation Day 2026 in Shanghai on October 29, 2026, as a proposed way to measure robot productivity in real-world settings.
- Worken combines active robots, autonomous working hours and an output efficiency ratio, rather than counting units built or delivered.
- The same event unveiled the AHA3 hardware platform, the GO-3 Preview model, WITA-Omni, the G2 Max industrial robot and Link-U OS 1.0.
Why counting robots is no longer enough
AGIBOT argues that manufacturing scale, technological capability and deployment milestones only tell part of the story. Two deployments involving similar numbers of robots can deliver very different levels of productivity, depending on how independently the machines operate, how often human intervention is required and how much acceptable work they complete.
That point connects to a wider debate in the AI economy. Other research published this week found that businesses are buying AI faster than they can turn it into real productivity gains, with gains shrinking as AI moves from individual tasks to whole organisations. Worken is AGIBOT’s attempt to apply the same discipline to physical machines: measure the output, not just the installation.
We want to move beyond measuring how many robots are built or delivered and begin measuring how effectively they can perform useful work, autonomously and consistently, in the real world.
That sentence comes from Peng Zhihui, AGIBOT’s Co-founder, President and CTO, who introduced Worken at the Shanghai event. He said the robotics industry has made remarkable progress in hardware, AI models and large-scale manufacturing, but now needs a more practical answer to how much productive work robots deliver in factories, warehouses and service environments.
How Worken is calculated
AGIBOT describes Worken as “Effective Intelligent Work Hours”, built from three factors multiplied together:
- Number of Active Robots: robots actively performing tasks in real-world environments, rather than units simply manufactured or held in inventory.
- Autonomous Working Hours: time spent working autonomously, rather than total powered-on time.
- Output Efficiency Ratio: productive output measured against a reference benchmark for a defined task, such as the amount of acceptable work completed by a human worker under established operating standards.
By connecting deployment scale, autonomy and task output, the framework aims to give a more complete picture of robotic productivity across applications such as industrial handling, logistics, inspection and commercial services. The concept also opens up a discussion about the economics of robotics: while hardware depreciates over time, improvements in software and AI models may enable the same machine to perform more useful work during its operating life, and measuring that output alongside lifecycle costs could help businesses judge the longer-term value of a deployment.
AGIBOT is clear that Worken is an evolving framework, not a finished standard. It says the framework will be explored and refined through real-world deployments and operational data, with task-specific benchmarks developed over time. The event also featured the release of the “2026 Embodied Intelligence Industry Innovation Report”, jointly compiled by China Economic Information Service and the Shanghai Industrial Innovation Center of the China Academy of Information and Communications Technology, which includes Worken and AGIBOT’s deployment examples in its industry analysis.
AHA3, GO-3 Preview and WITA-Omni
Worken was not the only announcement. AGIBOT unveiled AHA3, its third-generation humanoid hardware architecture and first commercial “Three Intelligences in One” hardware platform. AHA3 integrates the hardware foundations for locomotion, interaction and task execution intelligence into a unified architecture, built on advances in motor-drive technology, integrated computing and high-speed connectivity, including Wi-Fi 6 and 5G.
For sustained industrial deployment, AGIBOT also introduced a joint liquid-cooling system and specialized joints with IP67-rated protection, alongside a dexterous-hand family including OmniHand 3 Lite, OmniHand 3 and OmniHand 3 Ultra, plus the OmniPicker 3 tactile gripper.
In locomotion, the company highlighted advances in AGILE, its multimodal, end-to-end motion intelligence framework, including a newly introduced self-evolving agent capability designed to let robots explore the boundaries of their physical capabilities and improve through experience in real-world environments. AGIBOT describes this as a shift from motion control toward spatial mobility intelligence.
AGIBOT also debuted GO-3 Preview, a next-generation embodied AI foundation model built around a unified reasoning, prediction and action framework called WRAM (World Reasoning Action Model). The model is designed to help robots anticipate the consequences of their actions, adapt to changing environments and generalize beyond previously learned task sequences. According to evaluation results presented at the event, GO-3 Preview ranked first in task success rate and process score on RoboDojo, achieved the highest overall score on RoboTwin, and ranked first in overall score on RoboColiseum.
For human-robot interaction, AGIBOT showcased WITA-Omni, an omni-modal foundation model that integrates visual perception, speech, real-time dialogue, facial expressions and physical actions into a continuous interaction system. It supports full-duplex conversation, coordinated speech and movement, and on-device deployment, and AGIBOT says it ranks first on multiple benchmarks related to embodied interaction, including Daily-Omni.
G2 Max and Link-U OS 1.0 target real deployments
AGIBOT launched G2 Max, a heavy-duty industrial robot designed for manufacturing and logistics, with 21 degrees of freedom, a dual-arm handling capacity of up to 40 kg and an adjustable working height from 0.1 to 2.5 meters. Its force-controlled manipulation, omnidirectional mobility and safety features are designed for demanding material-handling tasks.
The company highlighted G2 family deployments including tablet production at Longcheer, semiconductor packaging and testing, CNC machine tending and logistics operations, where performance is assessed against cycle time, sustained operation and successful task completion. Beyond manufacturing, AGIBOT pointed to deployments in transportation and visitor-service environments, including Guangzhou Metro and the Chimelong tourism and entertainment complex, where robots guide visitors, answer questions and provide interactive services.
Finally, AGIBOT announced the official release of Link-U OS 1.0, which it positions as the industry’s first operating system purpose-built for embodied intelligence. Built on a unified hardware abstraction layer, an embodied intelligence-native kernel and an intelligent service framework, it connects robot hardware with upper-layer intelligent applications. The operating system has already been validated on mass-produced robots, is currently being trialled by selected ecosystem partners, and is planned to become fully open source by the end of 2026.
The announcements land at a moment when AI investment is under close scrutiny, from the AI chip boom lifting Samsung to a record quarterly profit to US regulators examining a major AI platform deal over competition concerns. AGIBOT’s pitch is that the next phase of that story will be judged not by announcements or unit counts, but by sustained, measurable work in the real world.
What to Watch Next
What to watch next is whether Worken moves from a proposed framework to a widely used reference. That will depend on the task-specific benchmarks AGIBOT says it will develop, and on operational data from the real deployments it highlighted in manufacturing, logistics, transportation and visitor services. Also worth watching is whether Link-U OS 1.0 becomes fully open source by the end of 2026 as planned, and whether the GO-3 Preview and WITA-Omni models hold their benchmark positions as they reach wider deployment.