BlockBeats News, May 23rd, According to official sources, DMind has released the first open-source Web3 vertical large model, DMind-1, surpassing Grok 3 and GPT-o3 in overall performance in the Web3 field. It is reported that DMind-1 is based on Alibaba's Qwen3-32B for fine-tuning, specifically optimized for the Web3 scenario, achieving state-of-the-art performance in benchmark evaluations, with an inference cost only one-tenth of mainstream large models. The lightweight version, DMind-1-mini, maintains high performance with faster response times and lower latency.
DMind stated in a tweet that general large models face issues such as low accuracy and high illusion in Web3 tasks. DMind-1 provides better inference capabilities and cost-effectiveness, suitable for the practical implementation of AI+Web3. DMind is an AGI open-source research organization jointly initiated by teams such as NFTGo and 2077ai, dedicated to building an open Web3 large model ecosystem, promoting technology open source, and developer collaboration.
免责声明：投资有风险，本文并非投资建议，以上内容不应被视为任何金融产品的购买或出售要约、建议或邀请，作者或其他用户的任何相关讨论、评论或帖子也不应被视为此类内容。本文仅供一般参考，不考虑您的个人投资目标、财务状况或需求。TTM对信息的准确性和完整性不承担任何责任或保证，投资者应自行研究并在投资前寻求专业建议。