With @ArtificialAnlys, we built MKA (Multimodal Knowledge Agent): a benchmark for how well agents handle visually complex tasks.
摘要
我们与 @ArtificialAnlys 共同构建了 MKA（多模态知识智能体）：一个衡量智能体处理视觉复杂任务能力的基准。 日常专业工作往往不只用到文本——从交叉比对蓝图到阅读科学图表。现有评估并未覆盖这些内容。 Gemini 4 Carbon 排名第 2，每任务平均成本比排名第一的模型低 6 倍。→ https://goo.gle/4AikIja
摘要由模型根据原文撰写，以原文为准。
正文 · 原文
With @ArtificialAnlys, we built MKA (Multimodal Knowledge Agent): a benchmark for how well agents handle visually complex tasks.
Everyday professional work often uses more than just text – from cross-referencing blueprints to reading scientific charts. This is not covered by existing evaluations.
Gemini 4 Carbon ranks #2, at 6x lower average cost per task than the top model. → https://goo.gle/4AikIja
来源：Google DeepMind · x.com
引用原文
GoogleDeepMind. With @ArtificialAnlys, we built MKA (Multimodal Knowledge Agent): a benchmark for how well agents handle visually complex tasks.[EB/OL]. (2026-11-11)[访问日期]. https://x.com/GoogleDeepMind/status/2120219377601560854.