PokemonGym is a server-client architecture platform designed for AI agents to be evaluated and trained in the Pokemon Red game. It provides game states via FastAPI, supports human-AI agent interaction, and helps researchers and developers test and improve AI solutions.

CSM 1B is a speech generation model based on the Llama architecture, capable of generating RVQ audio codes from text and audio input. The model is primarily used in speech synthesis and boasts high-quality speech generation capabilities. Its advantages include the ability to handle multi-speaker dialogue scenarios and generate natural and fluent speech through contextual information. This open-source model is intended to support research and educational purposes but is explicitly prohibited from being used for impersonation, fraud, or illegal activities.

CSM is a conversational speech generation model developed by Sesame. It can generate high-quality speech from text and audio input. The model is based on the Llama architecture and uses the Mimi audio encoder. It is mainly used for speech synthesis and interactive voice applications, such as voice assistants and educational tools. The main advantages of CSM are its ability to generate natural and fluent speech and its ability to optimize speech output through contextual information. The model is currently open-source and suitable for research and educational purposes.

WHAM (World and Human Action Model) is a generative model developed by Microsoft Research, specifically designed for generating game scenes and player behaviors. Trained on data from Ninja Theory's 'Bleeding Edge' game, it can generate coherent and diverse game visuals and controller actions. WHAM's primary advantage lies in its ability to capture the 3D structure of game environments and the temporal sequences of player actions, providing a powerful tool for game design and creative exploration. The model primarily targets academic research and the game development field, helping developers rapidly iterate on game designs.

GameFactory is an innovative general-purpose world model that focuses on learning from a limited amount of Minecraft gameplay video data and leverages prior knowledge from a pre-trained video diffusion model to generate new game content. Its core advantage lies in its open-domain generative ability, allowing it to create diverse game scenes and interactive experiences based on user-input text prompts and operational commands. It not only demonstrates strong scene generation capabilities but also achieves high-quality interactive video generation through a multi-stage training strategy and plug-in action control modules. This technology holds great promise in fields such as game development, virtual reality, and creative content generation. The pricing and commercial positioning are currently undefined.

Unbounded is an innovative infinite game that transcends the limitations of traditional finite, hard-coded systems by utilizing generative models. Inspired by James P. Carse's distinction between finite and infinite games, this game employs cutting-edge generative AI technology to create a character life simulation fully encapsulated by generative models. Drawing inspiration from sandbox life simulations, Unbounded allows players to interact with autonomous virtual characters using natural language commands, engaging in activities such as feeding, playing, and guiding characters, all facilitated by an open mechanism generated by large language models (LLMs), some of which may be emerging technologies. The development of Unbounded involved technological innovations in LLMs and visual generation, including a specialized, distilled large language model for real-time dynamic generation of game mechanics, narratives, and character interactions, as well as a new dynamic area image prompt adapter (IP-Adapter) to ensure visual consistency yet flexibility of characters across various environments.

PokemonGym is a server-client architecture platform designed for AI agents to be evaluated and trained in the Pokemon Red game. It provides game states via FastAPI, supports human-AI agent interaction, and helps researchers and developers test and improve AI solutions.

CSM 1B is a speech generation model based on the Llama architecture, capable of generating RVQ audio codes from text and audio input. The model is primarily used in speech synthesis and boasts high-quality speech generation capabilities. Its advantages include the ability to handle multi-speaker dialogue scenarios and generate natural and fluent speech through contextual information. This open-source model is intended to support research and educational purposes but is explicitly prohibited from being used for impersonation, fraud, or illegal activities.

CSM is a conversational speech generation model developed by Sesame. It can generate high-quality speech from text and audio input. The model is based on the Llama architecture and uses the Mimi audio encoder. It is mainly used for speech synthesis and interactive voice applications, such as voice assistants and educational tools. The main advantages of CSM are its ability to generate natural and fluent speech and its ability to optimize speech output through contextual information. The model is currently open-source and suitable for research and educational purposes.

Muse is a generative AI model developed by the Microsoft research team in collaboration with Xbox Games Studios, designed to support game ideation. Trained on a large-scale dataset of human gameplay, it can generate coherent game visual and action sequences. This technology showcases the potential of AI in game design, providing new creative approaches and experiences for future game development.

WHAM (World and Human Action Model) is a generative model developed by Microsoft Research, specifically designed for generating game scenes and player behaviors. Trained on data from Ninja Theory's 'Bleeding Edge' game, it can generate coherent and diverse game visuals and controller actions. WHAM's primary advantage lies in its ability to capture the 3D structure of game environments and the temporal sequences of player actions, providing a powerful tool for game design and creative exploration. The model primarily targets academic research and the game development field, helping developers rapidly iterate on game designs.

Pitch Lucy AI, developed by the UNDR Collective team, is an experimental crypto token game. Players pitch their favorite tokens to Lucy, an AI fund manager, in an attempt to convince her to invest. Powered by ZetaChain technology, this product provides users with a platform to interact with AI and challenge their abilities. It offers crypto token enthusiasts a chance to showcase their understanding of tokens and rewards players who successfully persuade Lucy, including earning a token discovery fee and having Lucy add the token to her investment portfolio. This innovative business model combines cryptocurrency investment with gamification, offering users a brand-new experience.

Secret Prompter is a fun game designed specifically for AI engineers, inspired by Wordle. Players must submit the best possible prompts within a limited number of tries to achieve the highest accuracy. The game updates daily, and player rankings are determined by accuracy and submission time. This product not only offers entertainment but also hones the prompt engineering skills of AI engineers, helping them better understand and optimize the input for AI models. The product is currently open for free, making it suitable for tech professionals and enthusiasts interested in AI.

WordPath is an intelligent game based on vocabulary associations, designed to challenge players' thinking abilities and language knowledge through enjoyable word connection tasks. It utilizes advanced language model technology to provide a rich variety of vocabulary connections, helping players expand their vocabulary and enhance cognitive flexibility. This product combines language learning with intellectual challenges in a fun game format, making it suitable for users of all ages to improve their language skills during leisure activities. Currently, the product is available as a website, allowing users to access and experience it anytime and anywhere without installation. It primarily targets language learners and those who enjoy intellectual challenges, offering both entertainment and educational value.

GameFactory is an innovative general-purpose world model that focuses on learning from a limited amount of Minecraft gameplay video data and leverages prior knowledge from a pre-trained video diffusion model to generate new game content. Its core advantage lies in its open-domain generative ability, allowing it to create diverse game scenes and interactive experiences based on user-input text prompts and operational commands. It not only demonstrates strong scene generation capabilities but also achieves high-quality interactive video generation through a multi-stage training strategy and plug-in action control modules. This technology holds great promise in fields such as game development, virtual reality, and creative content generation. The pricing and commercial positioning are currently undefined.