Education & Research Human-Machine Interaction Perception & Vision Software & Algorithm Provider
AGIBOT's WITA-Omni Preview Achieves Top Score in Audio-Visual Reasoning Benchmark
Original from RoboticsAndAutomationNews.com: AGIBOT’s foundation model tops benchmark test for audio-visual reasoning

AGIBOT's WITA-Omni Preview Achieves Top Score in Audio-Visual Reasoning Benchmark

AGIBOT announced that its WITA-Omni Preview multimodal foundation model has achieved the highest score on the Daily-Omni audio-visual reasoning benchmark, surpassing models from Alibaba, Google, and ByteDance. The model recorded an average accuracy of 85.21 percent, excelling in audio-visual alignment, event sequencing, and various evaluation metrics.

This achievement is significant as it demonstrates WITA-Omni Preview's advanced capabilities in understanding and reasoning across audio and visual information, which are crucial for embodied AI systems in dynamic environments. The benchmark evaluates models on their ability to associate sounds with visible events and understand the progression of situations, highlighting the importance of these skills in real-world applications.

Looking ahead, AGIBOT's WITA-Omni model is designed to enhance human-robot interactions by integrating movement and facial expressions with speech. The company has developed a human-centric multimodal interaction dataset to further improve the model's performance. No further timeline was disclosed at the time of publication.

Editor's Note

AGIBOT's advancements in audio-visual reasoning reflect a growing trend in the robotics industry towards more sophisticated AI systems capable of complex interactions. The integration of multimodal data is essential for enhancing the capabilities of robots in real-world applications, particularly in environments where understanding human behavior is critical. This development could influence future procurement decisions in robotics and AI technologies.

RobotToday Initiative

Robotics needs a service framework.

RSF defines a common language for robot service capability, lifecycle operations, certification pathways, and service-provider networks.

Share

Related Articles/News

AGIBOT's A3 Ultra Expands Deployment Across Hotels, Retail, and Metro Inspections

AGIBOT's A3 Ultra Expands Deployment Across Hotels, Retail, and Metro Inspections

AGIBOT announced on September 17 that its humanoid robot, A3 Ultra, is actively deployed in hotels, auto dealerships, supermarkets, and metro inspection scenarios. The company emphasizes the robot's ability to work autonomously, perform various tasks, and collaborate with other robots, although specific deployment numbers and performance metrics were not ...

Business Market China
Emerald AI, Google, and NVIDIA Form Alliance to Enhance AI Data Center Flexibility

Emerald AI, Google, and NVIDIA Form Alliance to Enhance AI Data Center Flexibility

Emerald AI, Google, and NVIDIA have announced the formation of the AI Energy Management Alliance (AEMA), aimed at advancing data centers that can dynamically manage electricity usage based on grid conditions. This initiative seeks to enhance AI infrastructure by enabling more efficient energy use, ultimately supporting community energy systems and reducin...

Moonshot AI Launches Kimi K2.8 Preview for Kimi Code and Work Users

Moonshot AI Launches Kimi K2.8 Preview for Kimi Code and Work Users

Moonshot AI has announced that the Kimi K2.8 Preview is now fully available for users of Kimi Code and Work. This update brings significant improvements, including coding and agent enhancements that are comparable to the flagship K3 model, adjustable thinking capabilities, and a context limit of 1 million across all tiers. The introduction of Kimi K2.8 Pr...

IMTS 2026 Preview: Record Orders and the Automation Story

IMTS 2026 Preview: Record Orders and the Automation Story

IMTS 2026 opens in Chicago with 1,788 exhibitors and record H1 machine tool orders. What the USMTO data says about automation, and the booths to see.

Market and Business News
AGIBOT Introduces AGILE 2.0 for Enhanced Perception in Humanoid Robots

AGIBOT Introduces AGILE 2.0 for Enhanced Perception in Humanoid Robots

AGIBOT has unveiled AGILE 2.0, a sophisticated framework designed for perception-driven locomotion and comprehensive control in humanoid robots. This new system allows for real-time modifications to movement and manipulation based on environmental feedback, significantly enhancing robot adaptability. The introduction of AGILE 2.0 is significant as it show...

robotics AI humanoid robots
Microsoft Releases September Security Updates for Windows and Google Launches Chrome 153

Microsoft Releases September Security Updates for Windows and Google Launches Chrome 153

On September 8, Microsoft initiated the distribution of its monthly security updates for all supported versions of Windows. This update addresses a total of 974 vulnerabilities, with 113 classified as critical. Among these, two vulnerabilities have already been exploited, prompting urgent updates for users. The significance of this update lies in its exte...

ROBOTTODAY WEEKLY September 7 – 11, 2026

ROBOTTODAY WEEKLY September 7 – 11, 2026

This week in robotics: JD.com plans to deploy three million robots across its logistics network, XPeng commissions its IRON humanoid production line, FANUC and Google launch an AI arc-welding agent, and Unitree shares fall 53% from their post-IPO high. September 7–11, 2026.

RobotToday Weekly
AGIBOT Launches Open-Source GE-Act 2.0, Competing with Unitree in Humanoid Robotics

AGIBOT Launches Open-Source GE-Act 2.0, Competing with Unitree in Humanoid Robotics

AGIBOT has officially released Genie Envisioner Act 2.0 (GE-Act 2.0), an open-source World-Action Model designed for training generalist manipulation policies. This release comes just after Unitree Robotics open-sourced its UnifoLM-WLA-1.0, marking a significant shift in the humanoid robotics landscape towards open-source models. The introduction of GE-Ac...

open-source China AGIBOT
AGIBOT Launches GE-Act 2.0, Advancing Robot Capabilities with New Model

AGIBOT Launches GE-Act 2.0, Advancing Robot Capabilities with New Model

AGIBOT has introduced GE-Act 2.0, a revolutionary native world-action model designed to validate a scalable pretraining path for embodied intelligence. This model demonstrates notable enhancements in robot performance as the volume of training data increases, showcasing its potential to transform robotic capabilities. The significance of GE-Act 2.0 lies i...

Robotics Machine Learning Artificial Intelligence
Google Partners with Mitti Labs for Major Rice-Methane Carbon Credit Agreement

Google Partners with Mitti Labs for Major Rice-Methane Carbon Credit Agreement

Google has entered into a significant agreement to purchase 1 million carbon credits from Indian startup Mitti Labs, marking the largest publicly announced deal for methane reduction in rice farming. This four-year contract will encompass rice farms across Karnataka, Andhra Pradesh, and Telangana, covering approximately 100,000 hectares at peak delivery. ...

Climate Startups Google
AgiBot Launches GE-Act 2.0 Native World-Action Model with Enhanced Data Scaling

AgiBot Launches GE-Act 2.0 Native World-Action Model with Enhanced Data Scaling

AgiBot has introduced GE-Act 2.0, a native world-action model that has been pretrained from random initialization using embodied data. This new model has significantly scaled from 300 to 30,000 hours of training, enabling it to perform zero-shot skills, including towel folding, across two different robot embodiments. The release of GE-Act 2.0 is significa...

Fanuc and Google Collaborate on AI-Driven Welding Robots for Enhanced Automation

Fanuc and Google Collaborate on AI-Driven Welding Robots for Enhanced Automation

Japanese industrial automation firm Fanuc has partnered with Google to create an AI-automated welding system. This innovative solution utilizes robots equipped with artificial intelligence that can interpret blueprints directly, thus removing the need for manual programming. Deliveries of this advanced system are expected to commence in December 2026. Thi...

Related Suppliers

Agibot 智元机器人

Major Chinese humanoid robot company founded in 2023, producing the A2 full-size humanoid at scale with $1B+ valuation and 10,000+ units shipped.

Humanoid

Google Ventures (GV)

Alphabet's independent venture capital arm investing in life sciences, AI, robotics, and enterprise technology

Scentience

US startup designing neuromorphic olfactory sensors, edge AI models, and developer APIs to give robots a sense of smell for navigation, security, and.

Beijing Ribri Technology Co., Ltd.

Chinese developer of opto-acoustic systems, long-range acoustic hailers, acoustic C-UAS detection and sonar payloads for unmanned and security platforms.

Maniformer 觅觅科技

AgiBot-incubated Chinese platform supplying embodied-AI training data, real and simulation datasets, and the MEgo data-collection hardware line.

Human Archive

Data lab capturing aligned multimodal human-interaction data to train robotics and embodied-AI foundation models.

Yaak Technologies GmbH

Berlin AI startup founded 2020, collecting multimodal driving data across 30+ German cities to train foundation models for autonomous vehicles and robotics.

Tommoro Robotics 투모로 로보틱스

Tommoro Robotics (Korea, f. 2022, SNU spinout) builds Robot Foundation Models and Physical AI software for humanoids and collaborative robots targeting.

Tera AI

San Francisco AI startup providing software-only spatial reasoning and visual navigation for robots, enabling GPS-free positioning using onboard cameras.

Software & Algorithm Provider

inJoin the RobotToday community on LinkedIn

Daily robotics news, in-depth analysis, conference highlights, and discussions with professionals worldwide.