# Google 發表 Gemini 3.5，最新一代結合前沿智慧與行動力的模型。

*genai · news · 2026-05-19 · Firstpost*

## Key points

- Google 發表 Gemini 3.5 Flash，迄今最快且最強大的代理與程式編碼模型。
- Gemini 3.5 Flash 現可在數秒內完成長期代理任務，成本約為以往的一半。
- 結合更新的 Antigravity 框架，Gemini 3.5 執行協作子代理以處理複雜工作流程。
- Gemini 3.5 Flash 是全球 Gemini 應用程式、搜尋 AI 模式及 Gemini Spark 助理的預設模型。
- Gemini 3.5 強化對網路及化學、生物、放射性及核子（CBRN）威脅的安全防護，降低有害內容生成風險。

At the Google I/O 2026 event on Tuesday, Google unveiled Gemini 3.5, its latest family of models that combines frontier intelligence with action. This marks a major leap forward in building more capable, intelligent agents. Google announced that it is kicking off the series with the new Gemini 3.5 Flash, which delivers frontier performance for agents and coding, excelling at long-horizon tasks and offering real-world utility. Gemini 3.5 Flash: Frontier performance for agents The new model offers intelligence that can rival large flagship models across multiple dimensions, while maintaining the speed expected from the Flash series. Google describes it as its strongest agentic and coding model yet, delivering frontier-level intelligence at exceptional speed—removing the need to trade quality for latency. Gemini 3.5 Flash: Agentic tasks at scale Google highlighted the model’s ability to handle long-horizon agentic tasks that previously took days or even weeks. These challenges can now be completed in seconds, often at around half the cost of other frontier models. It rapidly plans, builds, and iterates to solve real-world problems, whether developing applications, maintaining codebases, or assisting with financial documentation. Combined with the updated Antigravity framework, Gemini 3.5 becomes a powerful engine for executing collaborative sub-agents that tackle complex problems at scale. Under supervision, it can reliably execute multi-step workflows and coding tasks while maintaining frontier-level performance. Personal AI agents powered by Gemini 3.5 Gemini 3.5 Flash is now the default model for the Gemini app and AI Mode in Search globally, as announced at Google I/O. The new Gemini Spark uses Gemini 3.5 Flash and operates continuously, helping users manage their digital lives and take actions on their behalf under their direction. It has been rolled out to trusted testers, with a beta version coming to Google AI Ultra subscribers in the US next week. Gemini 3.5 has been developed in accordance with Google’s Frontier Safety Framework. Google has strengthened its cyber and CBRN safeguards, reducing the likelihood of harmful content generation. Gemini 3.5 Flash is generally available via Google Antigravity, the Gemini API in Google AI Studio and Android Studio, as well as Gemini Enterprise Agent and Gemini Enterprise.

**Companies:** Google
**Countries:** US

[Read the full story on Firstpost](https://www.firstpost.com/tech/google-i-o-2026-google-introduces-gemini-3-5-built-for-complex-agentic-workflows-14012994.html)

---

Canonical: https://newsio.io/zh-TW/n/6718cc69-d3c1-45ff-b492-8ffcb51ff775/google-gemini-3-5-flash
Summarized by Newsio from Firstpost. https://newsio.io/how-it-works