Google's Gemini 2.5 Flash-Lite is '一般に 利用できる' for developers looking for 急速な／放蕩な, cost-efficient AI
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 発表するd the "general availability" of its Gemini 2.5 Flash-Lite model, which is said to be its most cost-efficient and fastest 見解／翻訳／版 in the 2.5 series.
While developers can find $0.10 per 1M 記念品 input and $0.40 for 生産（高）, 2.5 Flash-Lite has already gone through some 激しい real-world シナリオs.
Companies like Satlyt have used the model to help 要約する telemetry data and 過程 衛星 data far quicker than before.
Gemini 2.5 Flash-Lite entered its preview 行う／開催する／段階 in June during the public availability of 2.5 Flash and 2.5 プロの／賛成の in the Gemini app, Vertex AI, and AI Studio.
The finale to Google's 最近の string of 2.5 models has arrived, as the company 最高潮の場面s a Gemini variant 以前 held in 実験（する）ing.
This afternoon (July 22), in an email to Android Central, Google 発表するd the "general availability" of Gemini 2.5 Flash-Lite, which has 出口d its preview 行う／開催する／段階. The company went into more in-depth 詳細（に述べる） in a Developers blog 地位，任命する, 明言する／公表するing this Gemini variant is its "fastest" and most "cost-efficient" AI model. The 地位，任命する 詳細（に述べる）s developers will find costs at $0.10 per 1M 記念品 input and $0.40 for 生産（高）.
力／強力にする is the 指名する of the game for Gemini 2.5 Flash-Lite. While 極端に lightweight, this Gemini variant still 保持するs Google's thinking 支配（する）／統制するs. More than that, the company 最高潮の場面s the model's "all-around high 質" 能力s in "coding, math, science, 推論する／理由ing, and multimodal understanding."
Google teases its Gemini 2.5 Flash-Lite model 能力s by 製図／抽選 attention to its 業績／成果 in a few real-world シナリオs.
From 衛星s to Data 分析
Companies like Satlyt have been using 2.5 Flash-Lite to help 減ずる latency for 批判的な onboard diagnostics for its decentralized space 計算するing 壇・綱領・公約. Gemini's fastest model has also helped 減少（する） Satlyt's 力／強力にする 消費 by ~30%, as it helps 過程 衛星 data and 申し込む／申し出 real-time summarization for telemetry data. Other companies like HeyGen have 報道によれば used Gemini 2.5 Flash-Lite to translate their ビデオ content in over 180 languages to reach 全世界の audiences.
DocsHound and Evertune have also spent time with the 2.5 Flash-Lite model to quicken long ビデオ 過程ing and 速度(を上げる） up 分析/報告（する）／憶測 世代.
Google 追加するs that developers can begin trying Gemini 2.5 Flash-Lite today (July 22) by typing “gemini-2.5-flash-lite” in their code. The AI model can be 設立する in Google's AI Studio and Vertex AI.
A Month in Preview
We've been on やめる a run with Google's 2.5 Gemini models over the past month. In June, the company brought its 2.5 プロの／賛成の and 2.5 Flash models into the public 注目する，もくろむ 経由で the Gemini app, Google AI Studio, and Vertex AI spaces. The most important part of that 告示 was 明らかにする／漏らすing that 2.5 Flash-Lite was entering its preview 行う／開催する／段階. This iteration （機の）カム only a few months after Google made the 2.0 見解／翻訳／版 public earlier this year.
At the time, Google 押し進めるd the cost-efficient and 急速な／放蕩な 味方するs of 2.5 Flash-Lite for developers looking to quicken their workload. In the event that you 要求する more thought 力／強力にする, there is Gemini 2.5 プロの／賛成の.