Unverified50% confidenceFactExact time

Gemini Robotics系列采用VLA（Vision-Language-Action，视觉-语言-动作）架构，将Gemini的多模态理解能力延伸到物理动作层面

1

Sources

50%

Confidence

Long-term

Relevance

8/31/2026

First Seen

Sources

Gemini Robotics 2深度解读：VLA架构与通用具身智能的未来

Related Entities

Gemini Roboticsmodel VLAconcept Geminimodel

Related Claims

Unverified本次演示将重点展示Gemini在个性化数据洞察、自动化重复性任务和内容创意生成三个维度的能力71% similar UnverifiedGemini Robotics 2相比上一代模型在泛化能力上有增强，能对未见过的物体、场景和任务进行更好的迁移和适应69% similar UnverifiedGemini的图像生成能力基于原生多模态架构，模型在训练阶段同时学习了文本和图像的表征69% similar UnverifiedGemini Robotics 2的目标是让一个统一的基础模型能够驱动多种形态的机器人完成多样化任务68% similar

Cite This Claim

Stable URI

https://kongchang.com/claim/829335

API

curl https://kongchang.com/api/v1/knowledge/claims/829335

MCP

get_claim(id=829335)