Google DeepMind 推出 SIMA 2：由 Gemini 驅動、能在 3D 虛擬世界中與你一同遊玩、推理與學習的 AI 代理
Original: SIMA 2: An Agent that Plays, Reasons, and Learns With You in Virtual 3D Worlds
Google DeepMind has officially introduced SIMA 2 (Scalable Instructable Multiworld Agent 2). Compared to its predecessor, the most…
Google DeepMind 推出新一代 AI 代理 SIMA 2，全面導入 Gemini 模型的能力。SIMA 2 不僅能在多個 3D 虛擬與遊戲世界中執行任務，還具備更強的推理、語言理解與即時決策能力。這項技術展示了 AI 如何從單純的指令接收者，演變為能在複雜互動環境中與人類協同合作的智慧夥伴。
Google DeepMind has officially introduced SIMA 2 (Scalable Instructable Multiworld Agent 2). Compared to its predecessor, the most significant transformation in SIMA 2 is its comprehensive adoption of Google's own Gemini multimodal model. This enables the AI agent not only to understand complex 3D virtual environments, but to possess deep capabilities for "thinking, understanding, and acting." SIMA 2 is designed to be a general-purpose companion that can play alongside human players, perform logical reasoning, and continue learning across a wide variety of interactive 3D virtual worlds — such as video games and simulation environments.
Free shows the 3-line summary; Pro unlocks the full deep summary (~300 words) so you never have to click through.
See Pro plans →Want the original English / full article?
Read on Google DeepMind Blog →Summaries are AI-generated; the original article is authoritative.