Web3 leader FLock.io has unveiled its milestone FLock Web3 Agent Model—a specialised LLM designed to execute complex tasks with precise function calls and on-chain analytics. Designed to power AI Agent projects with greater efficiency, and has already been snapped up by io.net, OpenGradient and HashKey Chain.
FLock Web3 Agent Model is more than just a blockchain assistant—it is a powerful tool driving AI automation across Web3. It is your very own Web3 expert, with accuracy outperforming industry-leading models GPT-4o, Gemini, DeepSeek and Qwen in benchmark evaluations.
While other models are general-purpose, FLock’s model is trained specifically for on-chain tasks. It understands blockchain logic natively, unlike traditional LLMs that require extensive prompting.
Web3 training resources and data are more scarce than their Web2 counterparts, limiting the performance of both proprietary and open-sourced models. FLock’s model solves this problem, bringing AI to Web3 natively rather than adapting Web2 AI models for a blockchain world.
Trained on high quality real-world and domain-specific Web3 data, and leveraging precise function-calling capabilities, this model can accurately interpret Web3 queries. It addresses key challenges:
- Understanding complex blockchain logic natively, without needing excessive fine-tuning.
- Interacting with smart contracts and dApps in real time, automating DeFi strategies, liquidity pool management, and multi-chain analytics.
- Executing precise function calls to streamline Web3 development, improve security, and reduce manual intervention in blockchain transactions.
FLock’s collaborations with io.net, OpenGradient, and HashKey Chain show how AI agents can deliver real-world value and become indispensable tools for on-chain intelligence, DeFi, and beyond.
This model can powerfully streamline Web3 workflows. That’s why io.net, a key partner, is integrating it with its newly launched IO Intelligence. This integration improves IO Intelligence’s agent performance and allows for more seamless API interactions. FLock benefits from io.net’s decentralised GPU network, making AI training faster, more accessible, and put to use in more real-world scenarios.
This isn’t the first time FLock has worked with io.net. Last August, they partnered to launch a first-ever Proof of AI.
OpenGradient’s deployment of FLock Web3 Agent Model exemplifies its suitability for real-time DeFi intelligence, on-chain analytics and Web3 queries.
HashKey Chain, too, is using FLock Web3 Agent Model exemplifies for on-chain financial solutions (DeFAI) projects and developer copilot tools. It simplifies interactions with APIs and delivers data-driven insights.
FLock Web3 Agent Model has yet more applications to accelerate AI projects. It excels at analysing liquidity pools, yield strategies and market trends; and cross-chain analytics, specifically to track, verify and interpret multi-chain transactions. It is also ready for smart contract querying and extracting insights directly from on-chain data.
“AI in Web3 has been held back by a lack of high-quality training data,” said FLock CEO Jiahao Sun. “The FLock Web3 Agent Model changes that.
“Built by and for the Web3 community on a decentralised platform, it’s designed to handle complex blockchain tasks with unmatched accuracy. We’re helping AI projects move from experimentation to impact.”
FLock Web3 Agent Model is the latest successful project from the FLock community. It was trained through a community-driven approach on AI Arena. Users participate in Kaggle-like competitions as task creators, training nodes, validators and delegators for FLOCK token incentives.
FLock.io is Web3’s AI powerhouse and home to the world’s first decentralised AI training platform. It is where developers, AI projects and data contributors unite with a common goal: to land AI’s place in the Web3 ecosystem and democratise AI development with the intersection of Federated Learning and blockchain.
With its Web3-native AI Agent Model, FLock is bridging the gap between blockchain and artificial intelligence. By offering unparalleled accuracy in on-chain analytics, smart contract execution, and decentralised automation, FLock Web3 Agent Model is redefining how AI interacts with blockchain ecosystems.
More about FLock
FLock.io is Web3’s AI-enabled ecosystem and home to the first decentralised AI training platform, combining Federated Learning and blockchain to revolutionise AI development. It enables secure, privacy-preserving training without centralising data, allowing communities to collaboratively create, train, and own AI models.
FLock’s ecosystem includes:
- AI Arena: A platform for competitive model training.
- FL Alliance: A space for privacy-focused collaboration to enhance models while preserving data sovereignty.
- Moonbase: A decentralised platform that hosts and refines AI models, rewarding contributors and driving the growth of a DeAI ecosystem.
The FLOCK token powers incentives, governance, and ecosystem growth, ensuring sustainable, decentralised innovation. Supported by top investors and recognized by the Ethereum Foundation, FLock is leading the democratisation of AI and transforming how AI interacts with blockchain.
Currently, over 30 models are running on FLock from both Web2 and Web3 use cases, with over 1,500 training and validation nodes joining from communities and institutions like Numerai, Kaggle, Foundry, and Tensorplex, contributing to over 19,000 models created.
FLock has raised capital across two major funding rounds: a $6 million seed round in March 2024 then a $3 million strategic round in December 2024. Investors include Digital Currency Group, Lightspeed Faction, and Animoca Brands.
