CtrlK
The AgiBot QUESTER1 (also known as Q1 or Prime Quester Q1) is a compact humanoid robot developed by AgiBot, a Chinese robotics company headquartered in Shanghai. Standing just 80 cm (2.6 feet) tall and weighing approximately 4 kg, the QUESTER1 is designed as a portable, developer-friendly platform for embodied AI research, education, and personal robotics experimentation. It was unveiled on December 31, 2025, by AgiBot co-founder and CTO Peng Zhihui via the Chinese video platform Bilibili.
Described as the world's first small-sized humanoid robot with full-body force control, the QUESTER1 represents AgiBot's entry into the personal and educational robotics segment. Unlike the company's larger A2 and G2 series, which target industrial and commercial applications, the Q1 is built to fit inside a standard backpack and is intended to lower the barrier to entry for students, independent developers, and startups working on embodied intelligence.
AgiBot Innovation (Shanghai) Technology Co., Ltd. was founded in February 2023 by two former Huawei engineers: Deng Taihua (CEO) and Peng Zhihui (CTO). Peng Zhihui, born in 1993, gained widespread recognition on the Chinese social media platform Bilibili under the handle "Zhihuijun" for his DIY robotics projects, including a robotic arm inspired by Iron Man and a self-driving bicycle. He was recruited into Huawei's Genius Youth Program in 2020, reportedly earning up to 2 million yuan (approximately US$287,000) per year, before departing in December 2022 to co-found AgiBot.[1][2]
Deng Taihua, the company's CEO, is a former vice president of Huawei who led the company's 5G campaign and helped build the Ascend AI ecosystem. The two co-founders share an alma mater in the University of Electronic Science and Technology of China (UESTC).
Since its founding, AgiBot has grown rapidly. The company shipped over 5,168 humanoid robots in 2025, ranking first globally by shipment volume according to analyst firm Omdia.[3] By March 2026, AgiBot had rolled out its 10,000th general-purpose robot.[4] The company's investors include BYD, Tencent, HongShan Capital (formerly Sequoia China), Hillhouse Investment, LG Electronics, and Mirae Asset, and it is reportedly planning a Hong Kong IPO in Q3 2026 targeting a valuation between US$5.1 billion and US$6.4 billion.[5]
The QUESTER1 sits within a broader product portfolio that spans multiple robot form factors:
| Product | Type | Height | Weight | Degrees of Freedom | Primary Application |
|---|---|---|---|---|---|
| Expedition A1 | Bipedal humanoid | 175 cm | 53 kg | 49 | Industrial automation |
| Yuanzheng A2 | Bipedal humanoid | 175 cm | 55 kg | 49+ | Commercial service, industrial |
| Expedition A3 | Bipedal humanoid | N/A | N/A | N/A | Entertainment, high-agility tasks |
| Lingxi X2 | Bipedal humanoid | 130 cm | 33.8 kg | 28 | Education, entertainment |
| Genie G1 | Wheeled humanoid | 130-180 cm | 150 kg | 26 | Industrial data collection |
| G2 | Wheeled industrial | Up to 180 cm | 185 kg | 26 | Precision manufacturing |
| QUESTER1 (Q1) | Bipedal humanoid | 80 cm | 4 kg | 28 | Research, education, personal |
| D1 | Quadruped | N/A | 3.6-9 kg | N/A | Patrol, inspection |
| OmniHand | Dexterous hand | N/A | N/A | 19 | End effector for manipulation |
The QUESTER1 complements rather than replaces the larger models. While the A2 series serves as AgiBot's flagship commercial humanoid and the G2 handles industrial-grade manipulation, the Q1 occupies a new niche: an affordable, crash-resistant platform that developers can use to test algorithms physically without risking expensive equipment.
Peng Zhihui formally revealed the QUESTER1 on December 31, 2025, in a video posted to Bilibili, the Chinese video-sharing platform where he had first gained fame for his robotics projects.[6] The announcement came at the end of a landmark year for AgiBot, during which the company reached the 5,000-unit production milestone and completed its acquisition of Swancor Advanced Materials, a Shanghai STAR Market-listed company.[7]
The reveal also followed a viral moment earlier in December 2025, when AgiBot's Lingxi X2 humanoid performed a dance routine alongside Chinese entertainer Wang Xingling, generating significant media attention in China and abroad.
Social media commentary noted Peng Zhihui's new title as chairman of Swancor (the publicly listed entity AgiBot acquired in July 2025 for approximately 2.1 billion yuan), and the Q1 launch was seen as a signal that AgiBot intended to expand beyond industrial and commercial markets into the personal robotics space.
The QUESTER1 is characterized by its compact dimensions and emphasis on full-body force control in a miniaturized form factor.
| Specification | Value |
|---|---|
| Height | 80 cm (2.6 ft / 31.5 in) |
| Dimensions (H x W x D) | 80 x 25 x 20 cm |
| Weight | 4 kg (8.8 lbs) |
| Locomotion | Bipedal |
| Walking Speed | 2.5 km/h |
| Degrees of Freedom | 28 |
| Arm Payload | 0.3 kg per arm |
| Deadlift Capacity | 1 kg |
| Specification | Value |
|---|---|
| Battery Type | Lithium-ion polymer |
| Battery Capacity | 10,000 mAh |
| Runtime | 2 hours per charge |
| Charging Time | 2 hours |
| Battery Lifespan | 3-5 years |
| Average Power Consumption | 50W |
| Peak Power Consumption | 100W |
| Specification | Value |
|---|---|
| Sensors | RGB cameras, IMU, gyroscope, force sensors, ultrasonic sensors |
| Navigation | Visual SLAM, indoor SLAM with obstacle avoidance |
| Connectivity | Wi-Fi, Bluetooth, USB, CAN bus, EtherCAT |
| Compatible Compute Platforms | NVIDIA Jetson, x86 PC |
| Operating Temperature | 0 to 40 degrees Celsius |
At 80 cm and 4 kg, the QUESTER1 is roughly one-eighth the weight and less than half the height of AgiBot's full-sized A2 humanoid (175 cm, 55 kg). Despite this dramatic reduction in size, the Q1 retains the same number of degrees of freedom (28) as the Lingxi X2, AgiBot's compact humanoid intended for entertainment and education. The Q1's miniaturized form factor makes experimentation safer and less costly, as crashes and falls involve far less kinetic energy than those of a 55 kg robot.
The defining technical innovation of the QUESTER1 is its miniaturized Quasi-Direct Drive (QDD) actuators. AgiBot reengineered its QDD joints to be smaller than an egg while retaining the precise force control and fast dynamic response characteristics found in its larger robots.[8] These joints are central to the Q1's full-body force control capability, allowing the robot to detect and respond to external forces across all its limbs simultaneously.
Quasi-Direct Drive actuators occupy a middle ground between direct-drive motors (which offer high force transparency but low torque density) and geared actuators (which offer high torque density but introduce friction and backlash). QDD systems use low-ratio gear reductions, typically below 10:1, to preserve much of the backdrivability and force sensitivity of direct-drive motors while achieving higher torque output. This makes them well suited for robots that need to interact safely with humans and environments, since the actuator can "give" when unexpected contact occurs rather than rigidly resisting it.
By shrinking these actuators to egg-sized proportions, AgiBot made it feasible to pack 28 force-controlled degrees of freedom into a robot small enough to carry in a backpack. The crash resistance of the Q1 stems partly from this joint design: because the QDD actuators are compliant, impacts are absorbed rather than transmitted rigidly through the structure.
The QUESTER1 runs on AgiBot's proprietary Agi-Soul AI platform, which provides the software layer for voice interaction, natural language processing, and learning capabilities.[9] Built-in functions include voice interaction, English language tutoring, guided dance coaching, and onboard positioning. The platform handles the integration between sensor inputs, language understanding, and motor control.
The Q1 is also compatible with ROS 2 and AgiBot's own AimRT middleware framework, providing flexibility for developers who want to use standard robotics tools or AgiBot's optimized alternatives.
The QUESTER1 supports teleoperation through cloud-connected servers, enabling remote control and data collection. This capability ties into AgiBot's broader strategy around data-driven robot learning: by collecting manipulation and locomotion data from Q1 units in the field, the company can expand its training datasets for foundation models like Genie Operator-1 (GO-1).
A central design philosophy behind the QUESTER1 is openness. AgiBot has positioned the Q1 as an open platform with multiple layers of accessibility:
Software Development Kit (SDK): A Python-based SDK provides programmatic access to the robot's sensors, actuators, and AI functions. The SDK supports integration with popular machine learning frameworks and robotics tools.
Hardware Development Kit (HDK): Hardware documentation and interfaces allow developers to modify or extend the robot's physical capabilities.
Open-Source Exterior Shell: AgiBot has released the Q1's exterior shell design files for 3D printing, allowing users to customize the robot's appearance without altering its internal mechanics.[10]
Zero-Code Motion Programming: A visual, block-based programming interface enables users without coding experience to program complex movement sequences. This lowers the barrier for educators, hobbyists, and researchers who want to explore embodied AI without writing low-level motor control code.
This open approach mirrors AgiBot's strategy with the X1, its larger open-source humanoid research platform. However, while the X1 targets university labs and research institutions with its US$20,000 price point and 130 cm frame, the Q1 targets a broader audience that includes individual developers, students, and makers.
The primary intended application for the QUESTER1 is as a physical research platform for embodied AI. Researchers can use the Q1 to test algorithms for locomotion, manipulation, reinforcement learning, and human-robot interaction in a real-world environment rather than relying solely on simulation. AgiBot has emphasized that the Q1's compact size and crash resistance shorten development cycles from months to days by allowing more aggressive and frequent physical testing.[11]
The Q1's 28 degrees of freedom and full-body force control make it suitable for research topics including:
The zero-code programming interface and built-in tutoring capabilities position the QUESTER1 as an educational tool. Its voice interaction and English tutoring functions suggest applications in language education and STEM learning environments where students can interact with a physical robot rather than a screen-based interface.
With its built-in voice interaction, dance coaching, and compact form factor, the Q1 also targets the emerging personal companion robot market. AgiBot envisions the Q1 bridging the gap between laboratory prototypes and consumer-grade personal devices, potentially serving as a stepping stone toward affordable household humanoid robots.
The QUESTER1's design reflects AgiBot's broader data-driven approach to robot intelligence. AgiBot World Colosseo, released in early 2025, is one of the largest open-source robot manipulation datasets in the world. The AgiBot World Beta dataset contains over 1 million trajectories spanning 217 tasks and 87 skills, with a total duration of 2,976.4 hours across more than 100 real-world scenarios.[12] The data was collected by over 100 homogeneous robots at a purpose-built 4,000+ square meter facility in Shanghai.
The AgiBot World paper was an IROS 2025 Best Paper Award Finalist and was subsequently published in IEEE Transactions on Robotics (TRO) in 2026.[13]
Building on this dataset, AgiBot developed Genie Operator-1 (GO-1), a generalist embodied foundation model that uses a Vision-Language-Latent-Action (ViLLA) framework. GO-1 predicts latent action tokens rather than directly outputting motor commands, bridging the gap between vision-language inputs and physical actions. In benchmarks, GO-1 achieved a 78% success rate on complex manipulation tasks, a 32-percentage-point improvement over prior state-of-the-art models.[14]
The Q1 serves as a potential data collection and policy deployment platform within this ecosystem. Because it shares the same software stack (Agi-Soul, AimRT, ROS 2 compatibility) as AgiBot's larger robots, policies trained on data from A2 or G2 platforms may be transferable to the Q1 with appropriate adaptation, and vice versa. The Q1's lower cost and smaller footprint could make it practical to deploy many units simultaneously for distributed data collection.
In July 2025, AgiBot released Lingqu OS, described as the world's first embodied intelligent operating system.[15] Lingqu OS provides a unified system layer between body hardware and intelligent applications. At the hardware level, it connects to components such as CPUs, GPUs, MCUs, sensors, drivers, and dexterous hands. At the software level, it supports core capability models for interaction intelligence, operation intelligence, and movement intelligence.
The framework includes a real-time middleware layer (AimRT), a standardized intelligent agent service layer, and a toolchain for simulation, training, and deployment. AgiBot began rolling out Lingqu OS under an open co-development model in Q4 2025, with the goal of establishing a unified system base and ecological interface for the entire embodied intelligence industry.
The QUESTER1 benefits from this infrastructure, as developers building on the Q1 can leverage the same software stack that powers AgiBot's larger commercial robots. This shared platform strategy means that research conducted on the Q1 can more easily translate to deployment on full-sized industrial platforms.
The QUESTER1 enters a growing market for compact and educational humanoid robots. Several companies have introduced miniaturized humanoid platforms targeting similar audiences:
| Robot | Company | Height | Weight | Degrees of Freedom | Target Market |
|---|---|---|---|---|---|
| QUESTER1 (Q1) | AgiBot | 80 cm | 4 kg | 28 | Research, education, personal |
| Reachy Mini | Hugging Face / Pollen Robotics | 35 cm | ~2 kg | 17 | Open-source research |
| AlphaMini | UBTech | 24.5 cm | 0.45 kg | 12 | Education, entertainment |
| Aelos | Leju Robotics | 36 cm | 1.8 kg | 19 | Education, hobby |
| NAO | SoftBank Robotics | 58 cm | 5.4 kg | 25 | Education, research |
The Q1 differentiates itself through several features: its 28 degrees of freedom exceed those of most competitors in the compact humanoid segment; its full-body force control via QDD joints is uncommon at this size class; and its integration with AgiBot's broader ecosystem (AgiBot World dataset, GO-1 foundation model, AimRT middleware) provides a pathway from small-scale experimentation to large-scale commercial deployment.
At a broader level, the QUESTER1 reflects a strategic trend across the humanoid robotics industry. Companies like Unitree Robotics, Figure AI, and Tesla have focused primarily on full-sized humanoids for industrial and commercial applications. AgiBot's Q1 suggests that the company sees value in cultivating a developer community and expanding the addressable market downward in both size and price.
The Chinese humanoid robot market is projected to grow significantly, with Chinese companies accounting for roughly 90% of global humanoid robot shipments in 2025.[16] China's humanoid robot output is expected to surge 94% in 2026, with Unitree and AgiBot together capturing nearly 80% of market share according to TrendForce.[17] The Q1 positions AgiBot to compete not only in the industrial segment but also in the consumer and educational tiers.
As of the December 2025 announcement, AgiBot did not disclose official pricing for the QUESTER1. Third-party sources have estimated a price range of US$5,000 to US$10,000, which would position it well below the X1 open-source platform (approximately US$20,000) and the Lingxi X2 (starting around US$14,000 for the Youth Edition).[18] Availability has been indicated for China and the United States, though specific distribution details and shipping timelines have not been formally announced.
AgiBot made its U.S. market debut at CES 2026 in January 2026, where the company showcased its full portfolio including the A2 series, X2 series, G2 series, and D1 quadrupeds. The QUESTER1 was part of the product range displayed at the event.[19]
The QUESTER1 represents a strategic expansion for AgiBot beyond its industrial and commercial roots. By creating an accessible, affordable humanoid platform with full-body force control, AgiBot aims to build a developer ecosystem analogous to the early personal computer community: a broad base of experimenters who can contribute to and benefit from advances in embodied intelligence.
The Q1 also fits into AgiBot's infrastructure-building strategy. The combination of Lingqu OS, AimRT middleware, the AgiBot World dataset, and the GO-1 foundation model creates a full stack that the Q1 can leverage, potentially reducing the time and expertise required to develop capable robot behaviors.
As humanoid robotics continues to attract investment and government support (particularly in China under the 15th Five-Year Plan for 2026 through 2030, which elevated robotics to a top-ten national industry track), compact and accessible platforms like the QUESTER1 may play an important role in expanding the community of robotics developers and accelerating the transition from laboratory research to real-world deployment.