澳大利亚推出主权 AI 模型 Agentaus 1:220B 参数,MMLU-Pro 得分 84.4%
Trellis Data 发布了澳大利亚自己的 220B 模型,MMLU-Pro 考了 84.4 分,GPU 用量只要同规模模型一半,数据全程不出澳洲。做主权 AI 或关注数据合规的人可以看看这个思路。
Trellis Data 发布澳大利亚本土 AI 模型 Agentaus 1,采用 220B 参数规模,在 MMLU-Pro 基准上取得 84.4% 的成绩,全球排名第 43 位。官方称该模型运行所需 GPU 数量仅为同规模模型的一半。
模型完全部署在澳大利亚境内的服务器上,数据不出境,用户输入不会被用于训练,删除对话后数据永久清除。产品目前提供免费起步方案,网址为 agentaus.com.au。
这则消息的背景是越来越多国家推进主权 AI 建设,由本国公司训练并在本土基础设施上运行模型。对澳大利亚政府机构和对数据驻留有合规要求的企业来说,Agentaus 1 提供了一个不依赖海外供应商的选择。
Bullish on sovereign AI. More countries are building their own AI. Australia's is Agentaus 1 from Trellis Data. It's a 220B model that scores 84.4% on MMLU-Pro and runs on half the GPUs of similar-sized models. It runs on Australian servers and never trains on your data. Michael Gately @MGately_au We built Australia’s own AI model, and today it’s open to the world. Agentaus 1 ( @AgentausTD ): 220B parameters. 84.4% on MMLU-Pro. Ranked 43rd in the world. It runs on Australian servers. Nothing ever leaves Australia, and nothing you type is used to train AI. Delete a chat and it’s gone for good. Most of the world’s AI answers to somebody else. This one answers only to you. Free to start: agentaus.com.au
This report is categorized under AITOP's "Models" desk. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows. Key entities and technical topics under continuous tracking include Agentaus 1, Trellis Data, MMLU-Pro, sovereign AI.
Source attribution: curated from elvis (@omarsar0).