The AgiBot WITA-Omni Full-Modal model achieved a high score of 85.21 in the DailyOmni benchmark. Furthermore, the model demonstrated supremacy in six out of eight metrics, surpassing systems such as Alibaba Qwen, Google Gemini, ByteDance Doubao, and NVIDIA.
The benchmark assesses the model's ability to simultaneously process audio and visual signals in real-world conditions. Among the tested parameters were audiovisual data consistency (86.13), temporal reasoning about events (84.64), complex cross-modal reasoning (85.06), as well as high results in tasks involving 30 and 60-second video comprehension.
Unlike general-purpose multimodal models optimized for interacting with digital content, WITA-Omni was originally developed for embodied interaction. This means that robots utilizing this model must understand, react to, and act within a physical environment in real time.
The innovation lies in the Thinker-Talker-Actor paradigm, which replaces the traditional sequential pipeline. Standard robot interaction processes perform speech recognition, semantic understanding, response generation, speech synthesis, and action output as separate, sequential stages, leading to accumulated latency. WITA-Omni, however, runs the Thinker multimodal reasoning core, the Talker speech streaming generation, and the Actor action and expression modules in parallel on a single timeline.
The Actor module takes as input not only the semantic state from the Thinker but also latent audio variables from the Talker. This allows gestures and facial expressions to synchronize with the rhythm of speech, pauses, accents, and emotional tone, eliminating discrepancies between what the robot says and how it moves. The architecture extends the Thinker-Talker paradigm, previously created for conversational AI, by adding the Actor module as a full output channel alongside speech.
The Thinker integrates textual, visual, audio, and video inputs through specialized encoders and lightweight projection layers into a unified representation space for decision-making regarding interaction and cross-modal reasoning. The Talker generates natural speech conditioned on the hidden states of the Thinker. The Actor controls the action head and expression head, using both the semantic state of the Thinker and the audio features of the Talker. The synchronization mechanism between the streams ensures that the robot's movements react to changes in speech speed, shifts in expression, and that physical actions maintain temporal coherence with verbal communication.
This result follows the achievement of the AgiBot world model Genie Envisioner-Sim 2.0, which previously took first place at WorldArena. These two achievements form a three-pillar structure of AI capabilities: WITA-Omni for interaction, the world model for understanding the physical world, and the robot hardware for acting within it. The victory in the ranking confirms that companies focused on embodied AI can achieve superior overall capabilities by optimizing for real-world constraints rather than adapting general-purpose models.
As robots transition from controlled factory environments to unstructured social spaces, the ability to synchronously process multiple sensory streams and respond with coordinated verbal and physical actions becomes a key differentiator. AgiBot defines synchronous multimodal interaction as a new competitive dimension in the embodied AI industry, separate from competition based purely on model quality dominating general AI.
ZTE has once again won the WAIC SAIL Star award thanks to its OEX supernode, which features a zero-cable design, the ability to connect 128 Graphics Processing Units (GPUs) in a single rack, and a comprehensive artificial intelligence infrastructure covering cloud supercomputers and Nubia agent phones.
The OEX supernode received the WAIC SAIL Star award for the second consecutive year. It solves the complex engineering challenge of connecting hundreds of GPUs at the scale of data centers while minimizing cable complexity, signal loss, and latency. Thanks to its orthogonal design, where computing racks and switching racks are connected vertically, the OEX supernode provides full interconnectivity for 128 GPUs in a standard rack without internal cables or a rear panel. This significantly reduces signal paths, lowers connection costs, and simplifies maintenance compared to traditional designs.
The architecture allows scaling up to 16,000 GPUs. ZTE developed proprietary chips for in-rack switching that simultaneously support the CLink protocol, overseen by China's Ministry of Industry and Information Technology (MIIT), and Broadcom's SUE protocol. This ensures GPU compatibility independent of the specific chip supplier, preventing vendor lock-in. The five-chip co-probing technology, which includes a Central Processing Unit (CPU), internal server interconnect, inter-rack communication, network card, and DPU, allows the system to dynamically select optimal configurations for different types of GPUs and workloads, making the OEX supernode a flexible platform adaptable to the rapidly evolving domestic GPU ecosystem.
ZTE's exhibition at WAIC was divided into four zones: AI Infrastructure, AI Applications, AI Smart Home, and AI Agent Smartphone. These zones presented a unified concept of AI—from computational power generation to end-user consumption. The infrastructure zone showcased the OEX supernode alongside the AIDC computing power cooperation architecture and the computing network structure.
The applications zone demonstrated the NewStart AIOS operating system, designed for industrial, home, and personal scenarios, as well as the Co-Claw agent corporate platform. This platform is used internally by ZTE to build an AI-managed organization and is replicated for the industrial, railway, mining, metallurgical, and energy sectors.
In the smart home zone, ZTE's products for home terminals were displayed, maintaining global market leadership for five consecutive years with over 100 million annual shipments, including 3 million AI-enabled devices. The agent smartphone zone featured the Nubia NaviX Ultra—the world's first production-ready agent smartphone, developed in collaboration with Doubao from ByteDance, which debuted at WAIC. Ni Fei, President of Nubia at ZTE, emphasized that the agent phone category represents the lower half of AI, performing tasks rather than just answering questions, which requires thorough engineering verification of all user pain points, native OS integration, stable manufacturing experience, and privacy assurance.
In addition to the exhibition, ZTE announced achievements in medical AI: generating medical reports with a tenfold increase in efficiency, creating HTA reports in 1–5 minutes, and assisting in cardiovascular disease diagnosis with 96% accuracy based on 10,000 clinical cases. Three humanoid robots were also presented: Xingzai, weighing 30 kg for commercial services; the heavy-duty Xinqing for industrial manufacturing; and the wheeled Xinwang. All operate on the Star Core automotive chip with the dexterous Xingqiao hands, achieving precision below 0.2 mm.
