FLock Web3 Agent Model accuracy surpasses GPT-4o and Gemini

[PRESS RELEASE – London, UK, March 12th, 2025]

Web3 leader FLock.io has unveiled its milestone FLock Web3 Agent Model—a specialized 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. Designed as a specialized Web3 expert, with accuracy outperforming industry-leading models GPT-4o, Gemini, DeepSeek, and Qwen.

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.

FLock is bringing AI to Web3 natively, rather than adapting Web2 AI models for a blockchain world. 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.

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 decentralized GPU network, making AI training faster, more accessible, and put to use in more real-world scenarios.

OpenGradient’s deployment of the 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 analyzing 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 decentralized 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 democratize 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 decentralized automation, the FLock Web3 Agent Model is redefining how AI interacts with blockchain ecosystems.

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 centralizing 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 decentralized 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, decentralized innovation. Supported by top investors and recognized by the Ethereum Foundation, FLock is leading the democratization 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 and then a $3 million strategic round in December 2024. Investors include Digital Currency Group, Lightspeed Faction, and Animoca Brands.

The post FLock Web3 Agent Model accuracy surpasses GPT-4o and Gemini appeared first on CryptoPotato.