Skip to main content

ASI-1 mini: The World’s First Web3 LLM, Designed for Agentic AI

ASI-1 Mini is the world’s first Web3-native large language model (LLM), engineered for autonomous, agentic workflows. It’s designed to think, adapt, and collaborate, ensuring your AI-driven workflows are smarter than ever. Developed by the Artificial Superintelligence Alliance, ASI-1 Mini is compact, fast, and built for real-world natural language applications. With integrated Knowledge Graphs, it delivers context-aware, adaptive interactions—ideal for powering AI agents, processing documents, and scaling decentralized systems.

Launch ASI1-miniDocsDeveloper Forum

Chat with ASI-1 mini

  • Knowledge Graph Integration
    ASI1-Mini structures your data into dynamic knowledge graphs — enhancing memory, personalization, and contextual accuracy.
  • Multiple Knowledge Graphs
    Manage and switch between different knowledge graphs for different domains, tasks, or users.
  • Image to Text
    Upload images and receive detailed, relevant language responses based on visual analysis.
  • Text to Image
    Describe what you want to see — ASI1-Mini generates original images on demand via integrated diffusion models.
  • Advanced Access with FET
    Hold FET to unlock priority access, premium features, and deeper integration options within the ASI1-Mini ecosystem.
Launch ASI1-mini

Agentic Intelligence, Adaptive Design

ASI-1 Mini blends multi-mode reasoning with a modular, agent-based architecture—built to think flexibly, collaborate autonomously, and scale intelligently.

  • Multi-Mode Reasoning: Switch between Multi-Step, Complete, Optimized, or Short modes for tailored depth and speed.

  • Mixture of Models (MoM): Dynamically selects specialized models to match the complexity of each task.

  • Mixture of Agents (MoA): Enables autonomous agent collaboration, enhancing scalability and distributed performance.

  • Context-Aware Adaptation: Adjusts reasoning pathways in real time for efficient, relevant, and goal-aligned outputs.

Classic Mode vs Knowledge Graph Mode

Switch easily between two modes:

  • Classic Mode: Stateless, fast, and familiar — traditional LLM behavior.
  • Knowledge Graph Mode: Stateful, structured, and evolving — built for long-term, personalized interaction.

Choose what works best for your workflow, or toggle between both.

Play with our API

ASI-1 Mini is a context-native API built for developers integrating intelligent, agent-ready language models into decentralized apps. Designed for modularity, reasoning, and data awareness.

  • Graph-aware input processing and retrieval
  • Agent tool-calling support (early stage)
  • Knowledge graph-enhanced context windows
  • Custom system prompts and memory options
  • JSON, Markdown, and plain-text output formats
  • Integration-ready for Web3 and dApp workflows
  • Usage metering and token-based access ($FET)
API DocumentationGet ASI1-mini API Key*

*login required

Technical Specifications

Feature Details
Type Lightweight Large Language Model
Knowledge Graphs Integrated support (multiple graphs per user)
Media Capabilities Image to Text, Text to Image
Personalization Memory-based preference retention
Document Support Custom data ingestion (PDFs, websites, structured text)
Integration Works with ASI Train, Create, and MeTTaCycle
Access Model Tiered access with premium features via FET holding
Deployment Available via web interface, APIs, and ASI platform tools

Already trusted by Web3 innovators and AI pioneers

Learn More
Learn More
Learn More

Start chatting with ASI1-mini

Ask questions, test prompts, or explore ideas with a context-native AI—powered by graph-enhanced reasoning and built for modular, agent-ready applications.