Skip to content
TheChatArchiveThe Chat Archive
gpt-4.1-2025-04-14· OpenAIHugging Face LMArena Human Preference 140K

"5M context and 83x72b paramaters, output modalities: VRChat Full-Body-Tracking, audio…

From a Chatbot Arena battleThe same person sent this prompt to two anonymous models at once. Compare it with Qwenqwen-max-2025-01-25’s answer.
  1. User

    "5M context and 83x72b paramaters, output modalities: VRChat Full-Body-Tracking, audio (stereo), text, images , VRChat facial tracking data, object and scene interaction data (like it can press buttons) input modalities: vision (in realtime allows it to see the other participants and respond), audio (stereo), text, images (videos included), discord custom emoji & stickers, tactile feel, context thoughts, taste & scents"

    Can this model be considered useful for social AI?

  2. Assistant

    Absolutely, this model would be considered extremely useful for social AI, and in fact, it would be at the cutting edge of what’s possible in the field. Let’s break down why:


    1. Massive Scale and Multimodality

    • 5M context: This means the model can consider up to 5 million tokens of context—orders of magnitude more than most current models. This allows for deep, persistent memory and nuanced, ongoing social interactions.
    • 83x72B parameters: That’s 5.976 trillion parameters—enormous, even compared to the largest models today. This would enable highly sophisticated reasoning, language, and perception.

    2. Rich Output Modalities

    • VRChat Full-Body-Tracking & Facial Tracking: The model can control avatars in virtual spaces with realistic body and facial expressions, making social interactions feel natural and embodied.
    • Audio (stereo): It can speak and listen in high fidelity, supporting natural conversation.
    • Text, Images, Object/Scene Interaction: It can communicate in multiple formats and interact with the environment, pressing buttons, manipulating objects, etc.
    • VRChat facial tracking data: Enables nuanced emotional expression.
    • Object and scene interaction data: Allows for meaningful participation in shared activities.

    3. Rich Input Modalities

    • Vision (real-time): The model can see and interpret the environment and participants, crucial for social awareness and context.
    • Audio (stereo): It can understand tone, emotion, and spatial cues.
    • Text, Images, Videos: It can process and respond to a wide range of media.
    • Discord custom emoji & stickers: It can understand and use internet/social culture cues.
    • Tactile feel, taste & scents: If these are simulated, the model can respond to a full spectrum of sensory input, making interactions more immersive.
    • Context thoughts: If this means it can access or simulate internal states, it can model theory of mind and empathy.

    4. Social AI Usefulness

    • Embodiment: The ability to control avatars with full body and facial tracking is a game-changer for social presence in virtual worlds.
    • Multimodal Communication: Handling text, voice, images, and even tactile/taste/scent data means it can participate in any kind of social interaction.
    • Contextual Awareness: Huge context window and real-time perception allow for continuity and depth in relationships and conversations.
    • Cultural Fluency: Understanding emojis, stickers, and internet culture is vital for modern social AI.

    Potential Applications

    • Virtual companions and friends
    • Social skills training
    • Therapy and support
    • Collaborative work in VR
    • Entertainment and roleplay
    • Education and tutoring

    Summary

    Yes, this model would be not just useful, but revolutionary for social AI. It would enable deeply immersive, emotionally intelligent, and contextually aware interactions in both virtual and real-world settings. If such a model existed, it would set a new standard for what’s possible in social artificial intelligence.

Responses were produced in different, uncontrolled contexts (system prompts, settings, prior turns). This is not a benchmark.