跳到正文
Gemini API:更新日誌·· 4 天前AI 評分74

Gemini API 開放 Genie 4 正式使用,並預覽 Tool retrieval

November 19, 2026

AI 導讀

Gemini API 已在 Interactions API(v1beta)正式推出 Genie 4,用於即時互動世界生成,支援透過 WebRTC 和 WebSockets 串流輸出可即時操控的 24 fps、16:9 影片及同步 48 kHz 立體聲音訊。

正文
  • Genie 4 (genie-4) generally available (GA): Released Genie 4 (genie-4) on the Interactions API (v1beta) for real-time interactive world generation, streaming steerable 24 fps 16:9 video with synchronized 48 kHz stereo audio over WebRTC and WebSockets. Seed worlds from text, a 16:9 reference image, or both, and steer live with WASD/arrow keys and mid-session text events. Available on the paid tier at pricing of $1.50 per 1M input tokens, $9.00 per 1M text output tokens, and $17.50 per 1M video output tokens (~$0.10/sec, stereo audio included); see Interactive world generation with Genie 4, the Genie 4 prompt and controls guide, and rate limits.

  • Tool retrieval (Preview): The Interactions API can now search your tool library and load function schemas on demand, so large tool sets no longer have to be sent in full on every request. Mark tools with defer_loading: true and add a {"type": "tool_search"} tool to let Gemini retrieve schemas for you, or set execution: "client" to run your own retrieval logic. Remote MCP servers support defer_loading as well.

    Tool Retrieval works with Gemini 3.6 and later models, including all Gemini 4 models, in the Interactions API. See the Tool retrieval guide.

來源:Gemini API:更新日誌 · ai.google.dev