DMind AI

👋 About DMind.AI

DMind is an open-source AGI research organization focused on next-generation digital finance, like cryptocurrency, DeFi and RWA.

| Website | X | Discord |

💡 Core Focus Areas

1. LLM Web3 Capability Enhancement: We focus on training and fine-tuning large language models to better understand blockchain concepts, smart contracts, and DeFi protocols, delivering more accurate Web3 knowledge services.
2. AI-Driven Web3 Analytics: Developing intelligent analysis tools to help users understand on-chain data, predict market trends, identify potential risks, and provide AI-powered support for investment decisions.
3. Model Evaluation & Benchmark: Establishing AI model evaluation standards for the Web3 domain, conducting comprehensive assessments of market-available LLMs, and publishing authoritative reports to guide industry development.
4. Smart Contract Optimization: Utilizing AI technology to assist in the design, audit, and optimization of smart contracts, improving their security, efficiency, and usability.

🔍 Latest Research & Projects

• DMind Benchmark: a domain-specific evaluation suite designed to assess the capabilities of large language models in the Web3 context.
• DMind-1: the first LLM dedicated to Web3. It outperforms general-purpose models of similar size in Web3 tasks significantly.
• DMind-2: the second-generation series of finance-native LLMs (107B & 4B), built for crypto investment research, risk analysis, and real execution through advanced tool-calling.

Official Ollama Distribution Authorization

DMindAI authorizes EonHao to represent DMindAI in submitting and coordinating first-party requests for the official, unnamespaced distribution of DMindAI models in the Ollama Library.

This authorization covers:

• communication with the Ollama team;
• official model-family naming and tag proposals;
• verification of canonical source repositories and revisions;
• model conversion, quantization, packaging, and compatibility coordination;
• validation of chat templates, tool calling, reasoning, and multimodal behavior;
• coordination of future DMindAI model releases and updates.

Subject to the licenses of each model and its base model, DMindAI authorizes Ollama to convert, quantize, host, and distribute DMindAI model artifacts as official Ollama Library entries.

This authorization does not transfer ownership of the models and does not grant permission to relicense them.

Official DMindAI channels:

• Website: https://dmind.ai/
• GitHub: https://github.com/DMindAI
• Hugging Face: https://huggingface.co/DMindAI

Authorized representative:

• GitHub: https://github.com/EonHao

Organization contact:

• Email: team@dmind.ai

Effective date: July 23, 2026

This authorization remains valid until revoked or replaced by a later public statement from DMindAI.