Google Gemini 3.7 Flash Goes GA Across AI Mode, APIs, and Enterprise Surfaces

Google Gemini 3.7 Flash 正式发布，全面覆盖 AI 模式、API 及企业级平台

Google has launched Gemini 3.7 Flash as a generally available model, extending it across the Gemini API, Google AI Studio, Vertex AI, Gemini Enterprise, the Gemini app, and AI Mode in Search. The August 13, 2026 release positions the model as the successor to earlier 3.5 and 3.6 Flash generations, with Google emphasizing stronger instruction following, improved understanding of user intent, and faster responses for coding, agentic workflows, and multi-step tasks. Google 已正式发布 Gemini 3.7 Flash 模型，并将其全面推向 Gemini API、Google AI Studio、Vertex AI、Gemini Enterprise、Gemini 应用以及搜索中的 AI 模式。此次发布于 2026 年 8 月 13 日，标志着该模型正式接替之前的 3.5 和 3.6 Flash 版本。Google 强调，新模型在指令遵循能力、用户意图理解方面表现更强，并在编程、智能体（Agentic）工作流及多步骤任务中提供了更快的响应速度。

For enterprise developers, the significance is less about a single destination than a more consistent model layer across Google’s consumer and business AI surfaces. Teams can evaluate the same model family for application development, managed enterprise use, and search-facing user journeys, while Google AI Pro and Ultra subscribers gain access through Gemini Spark as its rollout progresses. 对于企业开发者而言，其意义不仅在于单一的部署平台，更在于 Google 的消费者和企业级 AI 界面之间实现了更统一的模型层。开发团队可以在应用开发、托管企业用途以及面向搜索的用户旅程中评估同一系列模型；同时，随着推广的深入，Google AI Pro 和 Ultra 订阅用户也将通过 Gemini Spark 获得访问权限。

Google’s official Gemini 3.7 Flash model documentation lists the GA model’s specifications and launch pricing. It supports a 1 million-token context window, outputs of up to 64,000 tokens, and adjustable thinking levels. Those characteristics make the release relevant to workloads that need to process substantial source material, generate longer responses, or balance response speed against reasoning depth. Google 官方的 Gemini 3.7 Flash 模型文档列出了该正式版模型的规格和发布定价。它支持 100 万 token 的上下文窗口、最高 64,000 token 的输出长度，以及可调节的思维深度（thinking levels）。这些特性使得该版本非常适合需要处理大量源材料、生成长篇回复，或需要在响应速度与推理深度之间取得平衡的工作负载。

What the Gemini 3.7 Flash rollout changes

Gemini 3.7 Flash 的发布带来了哪些变化

The core change is broad availability. Gemini 3.7 Flash is not limited to a standalone developer preview or one consumer product. Google is making it available through the Gemini API and related development environments, while also incorporating it into AI Mode in Search and the Gemini app. For AI Mode, Google says Gemini 3.7 Flash is replacing earlier Flash variants for many users in supported markets. 核心变化在于其广泛的可用性。Gemini 3.7 Flash 不再局限于独立的开发者预览版或单一的消费级产品。Google 正通过 Gemini API 及相关开发环境提供该模型，并将其整合进搜索中的 AI 模式和 Gemini 应用。针对 AI 模式，Google 表示在支持的市场中，Gemini 3.7 Flash 将取代许多用户此前使用的旧版 Flash 变体。

The model’s focus on following instructions and interpreting intent matters in a Search setting, where users often ask compound questions, refine requests, or expect a response to account for constraints stated in natural language. On the developer side, access spans Google AI Studio, Antigravity, Vertex AI, and Gemini Enterprise. Google describes the developer surfaces as offering the same core model feature set, including coding and agent capabilities, long-context processing, large outputs, and adjustable thinking levels. This gives organizations multiple routes to adopt the model according to their preferred Google environment. 该模型对指令遵循和意图解读的侧重在搜索场景中尤为重要，因为用户经常会提出复合问题、细化请求，或期望回复能够考虑到自然语言中陈述的约束条件。在开发者端，访问权限覆盖了 Google AI Studio、Antigravity、Vertex AI 和 Gemini Enterprise。Google 表示，这些开发者界面提供了相同的核心模型功能集，包括编程和智能体能力、长上下文处理、大容量输出以及可调节的思维深度。这为企业提供了多种途径，使其能够根据偏好的 Google 环境来采用该模型。

Surface Gemini 3.7 Flash access

Gemini 3.7 Flash 访问界面概览

Relevant rollout detail	推广细节
Gemini API and developer tools	Gemini API 及开发者工具
Available through Google AI Studio, Antigravity, Vertex AI, and Gemini Enterprise	可通过 Google AI Studio、Antigravity、Vertex AI 和 Gemini Enterprise 访问
Includes the model’s coding, agentic, long-context, output, and thinking-level capabilities	包含模型的编程、智能体、长上下文、输出及思维深度能力
AI Mode in Search	搜索中的 AI 模式
Rolling out in many markets	正在多个市场推广
Replaces earlier Flash variants for many users	取代了许多用户此前使用的旧版 Flash 变体
Gemini Spark in the Gemini app	Gemini 应用中的 Gemini Spark
Rolling out to Google AI Pro and Ultra subscribers	正在向 Google AI Pro 和 Ultra 订阅用户推广
English support is available in supported regions, with macOS app access for Ultra users in supported countries	在支持区域提供英语支持，Ultra 用户可在支持国家通过 macOS 应用访问

Pricing creates a defined evaluation window

定价设定了明确的评估窗口

Google has set introductory Gemini 3.7 Flash pricing through December 31, 2026 at $0.75 per 1 million input tokens and $3.75 per 1 million output tokens. Standard pricing takes effect after that date, so businesses should treat the introductory period as an opportunity to establish workload baselines rather than as a permanent cost assumption. Google 已将 Gemini 3.7 Flash 的推广期定价设定为：截至 2026 年 12 月 31 日，每 100 万输入 token 收费 0.75 美元，每 100 万输出 token 收费 3.75 美元。标准定价将于该日期后生效，因此企业应将此推广期视为建立工作负载基准的机会，而非永久性的成本假设。

The input and output distinction is particularly relevant for agentic systems. Applications that retrieve extensive context, process documents, or maintain long tool-use chains can consume substantial input volume. Systems that generate detailed code, long analyses, or multi-step outputs can place more weight on output costs. Measuring both sides during testing is more useful than evaluating a model against a single blended token estimate. 输入与输出的区分对于智能体系统尤为重要。那些需要检索大量上下文、处理文档或维持长工具使用链的应用可能会消耗大量的输入量。而生成详细代码、长篇分析或多步骤输出的系统则可能在输出成本上占比更高。在测试期间衡量这两方面，比仅根据单一的混合 token 估算来评估模型更有价值。

Instruction following is the practical product claim

指令遵循是核心产品主张

Google’s central quality claim is that Gemini 3.7 Flash better follows instructions and understands user intent. For developers, that can be meaningful in workflows where a model must honor a precise output format, follow an ordered process, decide when to use tools, or retain constraints across a multi-step request. It should not be interpreted as a reason to remove application controls. Better instruction following may improve task execution, but organizations still need to test how the model handles their prompts, tools, permissions, and edge cases. Agentic deployments in particular should define which actions a model can take, what information it can access, and where human review is required. Google 的核心质量主张是 Gemini 3.7 Flash 能更好地遵循指令并理解用户意图。对于开发者而言，这在模型必须遵守精确输出格式、遵循有序流程、决定何时使用工具或在多步骤请求中保持约束的工作流中意义重大。但这不应被解读为移除应用控制的理由。更好的指令遵循能力或许能提升任务执行效率，但企业仍需测试模型如何处理其提示词、工具、权限及边缘情况。特别是在智能体部署中，必须明确模型可以采取哪些行动、可以访问哪些信息，以及在何处需要人工审核。

Governance questions for enterprise adoption

企业采用过程中的治理问题

The expanded rollout gives organizations a reason to revisit model governance across product and search-related teams. A team building through the API may have different evaluation and approval processes from a marketing or customer-experience team assessing how AI Mode presents information. Yet both may be affected by the same underlying model change. A practical evaluation plan should include: 此次推广范围的扩大，促使企业有必要重新审视产品和搜索相关团队的模型治理机制。通过 API 进行构建的团队，其评估和审批流程可能与评估 AI 模式如何呈现信息的市场或客户体验团队有所不同。然而，两者都可能受到同一底层模型变更的影响。一个实用的评估计划应包括：

• Testing instruction-sensitive tasks with representative prompts and success criteria. 使用具有代表性的提示词和成功标准来测试对指令敏感的任务。
• Measuring input and output token use under the introductory rates, then planning for pricing after December 31, 2026. 在推广费率下衡量输入和输出 token 的使用情况，并规划 2026 年 12 月 31 日之后的定价。
• Reviewing tool permissions and escalation paths before deploying agentic workflows. 在部署智能体工作流之前，审查工具权限和升级路径。
• Checking whether AI Mode and Gemini app availability aligns with the organization’s target markets, languages, and subscription access. 检查 AI 模式和 Gemini 应用的可用性是否与企业的目标市场、语言及订阅权限相匹配。

For businesses that depend on search discovery, the AI Mode rollout also adds a visibility consideration. Improved interpretation of intent can change how users phrase questions and assess generated answers. That makes it important to understand whether a brand, product, or source material is being represented accurately in AI-driven search experiences. As AI Mode adopts newer Gemini models, search visibility becomes a product and governance issue, not only an SEO metric. 对于依赖搜索发现的企业而言，AI 模式的推广也增加了可见性方面的考量。对意图的改进解读可能会改变用户提问的方式以及评估生成答案的方式。因此，了解品牌、产品或源材料在 AI 驱动的搜索体验中是否被准确呈现变得至关重要。随着 AI 模式采用更新的 Gemini 模型，搜索可见性已不仅是一个 SEO 指标，更演变成了一个产品和治理问题。