The Elder Scrolls VI remains shrouded in mystery since its announcement by Bethesda in 2018, but a recent statement from Xbox leadership provided new information about the project's status. Asha Sharma, the division's CEO, reported watching a demonstration of the RPG during a studio visit and positively highlighted its scope and narrative.
This statement was released on Tuesday, the 11th, after members of Xbox management visited Bethesda's facilities in Rockville, Maryland, in the United States. However, the event also saw demonstrations by workers protesting the company's recent layoffs.
In a social media post, Sharma mentioned having watched a 'live playthrough' of The Elder Scrolls VI, suggesting that a playable version of the RPG was shown during the visit. According to the executive, both the 'scale and grandeur are incredible,' and the story is 'even bigger.'
Despite these reports, the communication did not specify what development phase the project is in or which segments of the game were shown. It was also unclear whether Sharma played the presented version or merely observed someone else during the demonstration. However, the report indicates that Bethesda already has a functional version of parts of the anticipated RPG.
Another detail noticed by fans emerged from Sharma's own post: she used asterisks to mask what appeared to be a subtitle for The Elder Scrolls VI. This sequence of eight characters generated speculation in the community about whether the executive was hinting at the official title name. To date, Bethesda has not officially confirmed any subtitle for the next chapter of the franchise.
Since the announcement at E3 in 2018, the project has been known only as The Elder Scrolls VI. The anticipation for concrete data on the RPG has spanned almost a decade. The studio first presented The Elder Scrolls VI in 2018 through a brief teaser showing only a landscape and the series logo. Since then, the studio has directed much of its resources to other titles, notably Starfield, keeping information about the next Elder Scrolls scarce.
For this reason, any new information about the development attracts great attention from franchise followers. Despite the internal demonstration cited by Sharma, there is still no official release date set for the game. The post also gained prominence due to the context in which it occurred: while Xbox executives were visiting Bethesda's office, unionized workers were protesting in front of the building against the company's recent staff cuts.
Among the elements used in the protest was the 'Scabby' symbol, a giant inflatable rat traditionally used by American unions in labor demonstrations. According to disclosed data, approximately 1,600 employees were affected by the recent cuts at Xbox studios. New protests, organized by workers in collaboration with the Communications Workers of America (CWA), are scheduled for August 18 in various cities across the United States and Canada. Employees demand, among other issues, fairer contracts and safeguards against layoffs.
Although there is an internal restructuring at Xbox, Bethesda had previously stated that the planning for The Elder Scrolls VI was unaffected by the layoffs. Nevertheless, employees at studios linked to the company have expressed apprehension about the impact of the cuts on large game production. Currently, Xbox and Bethesda have not disclosed a release window, platforms, or more specific details about the gameplay and plot of The Elder Scrolls VI. The demonstration seen by Asha Sharma provides, at minimum, a new indication that the project is progressing, although the community is still awaiting a more detailed public presentation.
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.
