Gemini Robotics 2.5 logo

Gemini Robotics 2.5

Gemini Robotics 2.5: Google’s most capable robot model

掲載情報

HiCyou の掲載記録

掲載中
掲載先サイト
producthunt.com
掲載記録の作成日
2026/11/15
記録の最終更新
2026/11/15

掲載されている情報

  • 概要
  • 主な機能 5 件
  • ユースケース 3 件

このページは HiCyou の掲載記録を示すものであり、セキュリティ、所有権、または製品品質を認証するものではありません。

機密情報を共有したり購入したりする前に、外部サイトを確認してください。

Gemini Robotics 2.5 preview

Gemini Robotics 2.5とは

Gemini Robotics 2.5 is Google DeepMind's most capable vision-language-action model, translating what a robot sees and is instructed to do into hand and arm movements. According to DeepMind, it achieves a 30% to 100% higher success rate than the previous baseline and completes tasks 20% to 30% faster. In a Mercedes-Benz parts-kitting pilot with Apptronik's Apollo 2 humanoid, it reportedly reached 96% success. The model works across robot types and is rolling out to robotics partners.

主な機能

Vision-language-action model that converts visual input and instructions into robot hand and arm movements
Reported 30% to 100% higher task success rate than the previous baseline
Tasks completed 20% to 30% faster than the prior generation
Works across different robot types
Currently rolling out to robotics partners

ユースケース

  • Industrial automation tasks such as parts kitting on manufacturing lines
  • Programming humanoid robots to perform manipulation tasks from natural language and visual input
  • Applying a single robot model across multiple robot hardware types

スタートアップにこのツールが必要な理由

Robotics startups building on modern AI can leverage Gemini Robotics 2.5 to add vision-language-action capabilities to diverse robot hardware, potentially reducing the need to train task-specific models and enabling faster deployment of manipulation-focused automation.

よくある質問

Gemini Robotics 2.5の代替ツール

RT-2
NVIDIA Isaac GR00T
Physical Intelligence pi0