Newborn Town Partners with Viggle to Launch AI-Native Game: Enhancing Social Gaming Experiences with Next-Generation Physics-Native World Model

Recently, Newborn Town’s game-oriented social platform TopTop has launched “Slap Showdown”, an AI-powered 3D interactive mini-game developed in collaboration with Viggle, a company specializing in physics-native world models. The launch brings world model technology into social gaming, offering users a more personalized and engaging interactive experience.

Developed by TopTop and powered by Viggle’s self-developed Joint Space-Time (JST) world model, “Slap Showdown” lets users turn uploaded photos into customized 3D characters and take them straight into the game.

Launched in September, “Slap Showdown” has gained particular traction in Japan, among Spanish-speaking users in North America, and across the Middle East. It is now available in multiple countries and regions worldwide.

From Better Gameplay to a Richer Ecosystem: How World Models Could Redefine Social Gaming on TopTop

In terms of gameplay, Slap Showdown combines AI-generated 3D characters with multiplayer interaction. Players can simply upload a photo from their phone and generate a personalized 3D character in around 30 seconds, whether it’s an anime protagonist, a cyberpunk cyborg, or a LEGO minifigure. They can then bring their custom characters into the game and face off against other players in fun, interactive battles.

As players use controls such as press-and-hold, tap, and drag, the 3D characters respond with realistic, physics-based movements. This takes AI-generated characters beyond visual customization, integrating them directly into the gameplay experience.

According to TopTop’s operations team, “Slap Showdown” has attracted a growing base of highly engaged, high-spending-potential players in its first month since launch. The game has recorded a 97% match completion rate, while its game-start rate has also exceeded the platform-wide average.

The depth of user engagement is further illustrated by individual player activity. One player in Japan completed 1,041 matches over 22 days, reaching a daily peak of 151 matches and averaging one match every 9.5 minutes. Another particularly avid player in Eastern Europe completed 735 matches in just 11 days, averaging 66.8 matches per day and peaking at 126 matches in a single day.

For a UGC-driven social gaming platform such as TopTop, mini-games serve as entry points and traffic drivers, while social features underpin retention and monetization. Personalized 3D avatars add another dimension to that experience. Beyond their visual appeal, they give users new ways to express themselves, interact with friends and share memorable moments. In turn, richer gameplay and more meaningful social interactions can help improve retention, attract new users and strengthen engagement across the platform.

Slap Showdown marks a successful step in TopTop’s efforts to deepen its application of AI technology, as well as a starting point for using world models to enhance gaming and social experiences. TopTop’s operations team said future game releases could make greater use of AI-powered 3D generation, extending beyond avatars to include customized environments. Existing games could also be revitalized by the introduction of new gameplay features, potentially driving stronger retention and user acquisition.

Beyond enriching gameplay, world model capabilities could also reshape social interaction itself, enabling more immersive, face-to-face experiences in virtual spaces. “We’re constantly exploring how AI can transform gaming and social interaction, while bringing new energy and capabilities to our UGC ecosystem,” TopTop’s operations team said.

Viggle Unveils Physics-Native Model JST-2, Setting Its Sights on Becoming the Infrastructure Layer for Physical AI

The world model technology powering “Slap Showdown” is provided by Viggle.

Founded in 2022, Viggle has focused on developing physics-native world foundation models. Its self-developed JST model can generate 3D characters, motion and environments directly, providing a foundation for gaming and a growing range of AI applications.

Viggle previously attracted widespread attention with Meme Maker, its AI-powered video creation product. The tool enables users to animate static characters and generate videos by applying specified movements, making AI character animation and video creation more accessible. In August 2024, Viggle raised USD 19 million in a Series A funding round led by Silicon Valley venture capital firm Andreessen Horowitz (a16z).

However, video creation was an early application of this technology. Viggle’s longer-term ambition is to make JST a foundational technology for AI-powered experiences across gaming, social platforms and other applications.

Viggle is now further expanding its technological boundaries. On 19 October, Viggle launched JST-2 in full, likening its significance for world models to GPT-2’s role in the evolution of language models.

Unlike many world models that infer physical structure indirectly from pixels, JST-2 is built around a physics-native architecture designed to learn the structure and dynamics of the 3D world directly. Its underlying representation consists of units resembling digital “atoms.” Each has a defined position in three-dimensional space, along with physical properties such as material, color and mass, and information describing how it moves and responds to forces. This allows the model to do more than generate convincing images. It can represent and generate 3D characters, motion and scenes with an underlying understanding of their physical properties.

To support this approach, Viggle has developed an integrated technology stack spanning tokenization, model architecture, data engine, pre-training and post-training. After four years of research and development, JST-2 can generate characters, motion and scenes within a single unified architecture.

Compared with other approaches, JST-2 also demonstrates cost efficiency. JST-2 can generate a 3D scene in seconds, at an inference cost of less than one US cent. Once created, the content can be rendered locally without repeatedly running the foundation model to generate each frame.

It is noteworthy that world model technology remains at an early stage, but Viggle is already working toward more advanced capabilities.

Its next-generation model, JST-3, is intended to become a comprehensive world simulator. Such a system could underpin future game and graphics engines, while also serving as a virtual training environment for robotics and other forms of physical AI.

The longer-term goal is for JST-3 to learn the physical rules governing its simulated world directly from data, rather than relying on rules explicitly written by developers.

For Newborn Town, its investment in and partnership with Viggle represents an important step in exploring how advanced AI can reshape social gaming, promoting a deep integration of AI technology with its business operations.

As a leading global social entertainment company, Newborn Town has continued to deepen the use of AI across key areas in recent years, including product R&D, social recommendation systems, intelligent operations, safety and risk management, and marketing and user acquisition, supporting ongoing business growth.

Meanwhile, the company also continued to invest across the AI ecosystem, backing projects in areas including world models, AI companionship, AI Agents and AI interactive gaming. These investments have further enriched its AI application ecosystem, with Viggle among the AI companies previously backed by Newborn Town.

“Slap Showdown” represents an early step in bringing world models into social gaming. For Newborn Town and Viggle, the next challenge is to turn increasingly sophisticated AI capabilities into experiences that users find genuinely engaging — and compelling enough to keep coming back.

*For more information about Viggle’s JST-2 world model, visit https://viggle.ai/jst