Google宣布推出Gemini 3.8 Flash，这是该公司在六周内发布的第三个Flash模型变体1。新模型包括标准版和Cyber版两个版本，其中标准版面向智能体任务和软件开发，Cyber版针对漏洞检测和防御进行了优化1。
根据Google的说法，Gemini 3.8 Flash在复杂任务上执行更多推理步骤，相比前代Gemini 3.7 Flash"工作更努力"2。该模型新增了迭代工具调用能力2。开发者可通过API以优惠价格访问这一模型，年底前输入令牌价格为每百万0.75美元，输出为每百万3.75美元12。
不过Google警告称，模型在提升性能时可能会消耗更多令牌，特别是在更高的努力水平下，因此实际使用成本可能高于入门价格2。开发者可选择继续使用Gemini 3.7 Flash以最小化令牌使用2。
Google has announced the launch of Gemini 3.8 Flash, marking the company's third Flash model release within six weeks.1 The new model comes in two variants: a standard version designed for agentic tasks and software development, and a Cyber edition optimized for vulnerability detection and defense.1
The model performs more reasoning steps on complex tasks compared to its predecessor Gemini 3.7 Flash and features iterative tool-calling capabilities.2 Google has cautioned that the upgraded performance may come at a cost, warning that "the model might use more tokens to maximize performance, particularly at higher effort levels," potentially resulting in higher actual usage costs despite maintaining the same introductory pricing structure.2 Developers retain the option to continue using Gemini 3.7 Flash to minimize token consumption.2
Pricing for Gemini 3.8 Flash has been set at $0.75 per million input tokens and $3.75 per million output tokens through the end of the year, representing a discounted rate compared to the standard pricing of $1.50 and $7.50 respectively.1 Developers can access the model through API at these promotional rates.1
评论
还没有评论，欢迎留下第一条。