Databricks 发布 Genie:企业级自主 AI 同事,任务准确率达 91%
Databricks 新发的 Genie 在企业任务上跑出 91% 准确率,token 用量不到对手一半,靠的是从自家数据学习业务逻辑的 Ontology 层。做企业数据或智能体落地的同学值得对比一下这条路线。
Databricks 于 2026 年 2 月 26 日发布新一代 Genie,定位为常驻型自主 AI 工作伙伴,在企业任务基准上取得 91% 的准确率,是两个前沿智能体中更强者的 1.6 倍,同时 token 消耗低于对手的 50%。
Genie 的核心是 Genie Ontology 上下文层:它自动从企业自身的数据、查询、仪表板和 Notebook 中学习业务实际运作方式,把企业的语义结构沉淀为智能体可直接调用的上下文,而非依赖通用模型的理解能力。
这与多数直接在通用模型外层做编排的企业级智能体形成差异:Databricks 主张性能瓶颈不在模型智力而在上下文供给,企业数据平台恰恰是上下文最富集的位置,这一定位直接影响企业采购智能体方案时的技术路线选择。
Ali often says, “AI doesn’t have an intelligence problem. AI has a context problem.” I 100% agree.
Today, Databricks is launching a new Genie: an always-on, autonomous AI coworker.
Genie scores 91% on enterprise tasks, 1.6X the accuracy of the stronger of 2 frontier agents, using <50% of the tokens.
Our edge is Genie Ontology, a context layer that automatically learns how a business actually works from its data, queries, dashboards, and notebooks.
Use Genie in your business!
This report is categorized under AITOP's "Products" desk, focusing on US frontier AI labs and Silicon Valley technology leaders. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows and US-China foundation model benchmarks and commercialization comparison. Key entities and technical topics under continuous tracking include Databricks, Genie, 智能体, 企业数据平台.
Source attribution: curated from Yuchen Jin (@Yuchenj_UW).