Google's 最新の AI model, Gemini 2.0, is here to 勧める in the 'age ntic 時代'
When you buy through links on our articles, 未来 and its syndication partners may earn a （売買）手数料，委託（する）／委員会／権限.
What you need to know
Google is 解放（する）ing Gemini 2.0 starting today, which will 力／強力にする what it calls the "agentic 時代."
Today, Gemini 2.0 Flash 実験の is 利用できる to developers in the Google AI Studio and Vertex AI using the Gemini API.
Gemini 2.0 Flash 実験の is also 利用できる in a 雑談（する）-optimized form in the Gemini web （弁護士の）依頼人 starting today, and the 十分な 顔触れ of Gemini 2.0 models will make their way to more Google 製品s and services next year.
概略で 10 months after Google's 現在の 世代 of AI models were 解放（する）d to the public, the company is previewing the 未来: Gemini 2.0. Starting today, the smaller Gemini 2.0 Flash 実験の model will be 利用できる to developers and Gemini 使用者s, with more sizes and 実施s to come next year.
Google 見解（をとる）s the next step 今後 for AI as the "agentic 時代," and in a blog 地位，任命する, the company 輪郭（を描く）d how Gemini 2.0 models are 目的-built to 力／強力にする AI スパイ／執行官s. This will be the 創立／基礎 for Google's most ambitious 事業／計画（する）s, from multimodal helpers to Chrome 拡張s that can do your browsing for you.
"Over the last year, we have been 投資するing in developing more agentic models, meaning they can understand more about the world around you, think 多重の steps ahead, and take 活動／戦闘 on your に代わって, with your 監督," wrote Google CEO Sundar Pichai. "With new 前進するs in multimodality ― like native image and 音声部の 生産（高） ― and native 道具 use, it will enable us to build new AI スパイ／執行官s that bring us closer to our 見通し of a 全世界の／万国共通の assistant."
Although only an 実験の model of Google's smallest AI model is 利用できる today, there are a few 重要な 推論する／理由s to be excited. 顕著に, Google's Gemini 2.0 Flash model not only outperforms Gemini 1.5 Flash, but also the larger and more powerful Gemini 1.5 プロの／賛成の model in 確かな （判断の）基準s. Gemini 2.0 Flash now supports both multimodal inputs and 生産（高）s, meaning the model can natively 生成する images, text, speech, or a combination of the three.
Gemini 2.0 Flash 実験の is 利用できる globally as an 選択 in the Gemini web （弁護士の）依頼人 today, and will come to the Gemini 動きやすい app soon. Additionally, developers can try out the 実験の 見解／翻訳／版 of Gemini 2.0 Flash using the Gemini API in Google AI Studio or Vertex AI.
結局, Gemini 2.0 will 力／強力にする most or all of Google's AI-equipped features. The models are 特に tailored for use with AI スパイ／執行官s, like 事業／計画（する） Astra, which was previewed at Google I/O 2024. It's a multimodal AI helper that can take in what's happening around you and answer questions with the 状況 of your 環境. There are also new 事業／計画（する）s, like 事業／計画（する） 水夫. 現在／一般に, 事業／計画（する） 水夫 is a Chrome 拡張 存在 実験（する）d as a 研究 原型 that can 扱う your browsing needs for you.
Google's agentic 見通し is ambitious, with plenty more 使用／適用s. There's Jules, a helper made for developers that 統合するs with GitHub workflows. Another proof-of-概念 is the company's 共同 with Supercell, which 実験（する）s how AI スパイ／執行官s can be 適用するd to 戦略 and 模擬実験／偽ること games.
Finally, there's now a new 深い 研究 方式 in Gemini 前進するd that uses long 状況 windows and 前進するd 推論する／理由ing to 機能（する）／行事 as a 研究 assistant.
Many of Google's Gemini 2.0-力／強力にするd ideas aren't 利用できる now, but are in active 開発. 事業／計画（する） Astra, for example, is 存在 実験（する）d externally through Google's 信用d tester program. However, some are 利用できる starting today, like Gemini 2.0 Flash 実験の in Gemini and Gemini 2.0 in AI Overviews ― which is in 限られた／立憲的な 実験（する）ing and will be more 広範囲にわたって 利用できる next year.