Google announces Gemini 3.7 Flash just three weeks after previous release
Google announces Gemini 3.7 Flash just three weeks after previous release
谷歌在上一版本发布仅三周后即宣布推出 Gemini 3.7 Flash
Google is announcing a new Gemini model today, but it’s not the long-awaited 3.5 Pro. Gemini 3.7 Flash is now rolling out to replace 3.6 Flash, which itself was released only three weeks ago. This new “workhorse” model is supposedly the product of core optimizations and developer feedback, offering improved coding and agentic performance. And Google is hoping to counter the lower cost of some competing models with a lower “introductory price” for 3.7 Flash.
谷歌今天宣布推出一款新的 Gemini 模型，但它并不是人们期待已久的 3.5 Pro。Gemini 3.7 Flash 现已开始推送，旨在取代三周前才发布的 3.6 Flash。这款新的“主力”模型据称是核心优化和开发者反馈的产物，在编程和智能体（agentic）性能方面有所提升。谷歌希望通过为 3.7 Flash 提供更低的“入门价格”，来应对竞争对手模型带来的成本压力。
According to Senior Director Tulsee Doshi, Gemini 3.7 Flash is noticeably better at coding than the previous Flash release. She cites a jump in the FrontierCode 1.1 Main test from 34.4 to 43.6 percent and DeepSWE v1.1 going from 49 to 65.3 percent. As for the vibes, Gemini 3.7 Flash’s WebDev Arena score has risen to 1,588 from 1,538.
据高级总监 Tulsee Doshi 称，Gemini 3.7 Flash 的编程能力明显优于之前的 Flash 版本。她指出，该模型在 FrontierCode 1.1 Main 测试中的得分从 34.4% 跃升至 43.6%，在 DeepSWE v1.1 测试中从 49% 提升至 65.3%。在综合体验方面，Gemini 3.7 Flash 的 WebDev Arena 得分也从 1,538 分上升至 1,588 分。
People turning to Gemini and hoping it will “know” things may also see modest improvements in Gemini 3.7 Flash. The GDP.pdf benchmark, which measures how well a model can process complex documents, has gone up to 34 percent versus 22 percent with 3.6 Flash. AutomationBench tests how well models can execute common business workflows, and Gemini 3.7 Flash rose to 30.4 percent from 3.6’s 17 percent score.
对于那些希望 Gemini 能“知晓”更多信息的用户来说，Gemini 3.7 Flash 也带来了一些适度的改进。用于衡量模型处理复杂文档能力的 GDP.pdf 基准测试得分从 3.6 Flash 的 22% 提升到了 34%。在测试模型执行常见业务工作流能力的 AutomationBench 中，Gemini 3.7 Flash 的得分也从 3.6 版本的 17% 上升到了 30.4%。
Those numbers certainly are higher. But are they sufficiently different to support a new model release just three weeks after the last one? This may be more about maintaining the appearance of constant improvements in Google’s AI. Throughout 2024 and 2025, Google rapidly made up ground to rival the best AI coming out of competing AI labs. Things appear to have slowed in 2026, though. At I/O in May, Google promised that the flagship Gemini 3.5 Pro would launch in June, but that never happened. Developers and businesses that have invested in Google’s AI tools will have to make do with a slightly better Flash model for now.
这些数字确实有所提高。但它们是否足以支撑在上一版本发布仅三周后就推出新模型呢？这可能更多是为了维持谷歌 AI 持续进步的形象。在 2024 年和 2025 年期间，谷歌迅速追赶，以抗衡竞争对手 AI 实验室推出的顶尖 AI 产品。然而，到了 2026 年，进展似乎有所放缓。在 5 月的 I/O 大会上，谷歌曾承诺旗舰版 Gemini 3.5 Pro 将于 6 月发布，但最终并未兑现。目前，投资了谷歌 AI 工具的开发者和企业只能先凑合使用这个稍好一点的 Flash 模型。
Google is hoping a lower price will keep developers engaged with Gemini. Gemini 3.7 Flash will be available through the end of the year at a rate of $0.75/1M input tokens and $3.75/1M output tokens, which is half of what 3.6 Flash costs. OpenAI recently dropped the price of its GPT 5.6 models, with the Flash-like Luna version at $0.20/1M input tokens and $1.20/1M output tokens.
谷歌希望通过更低的价格来保持开发者对 Gemini 的使用粘性。Gemini 3.7 Flash 在今年年底前的定价为每百万输入 Token 0.75 美元，每百万输出 Token 3.75 美元，这仅是 3.6 Flash 价格的一半。OpenAI 最近下调了其 GPT 5.6 模型的价格，其中类似 Flash 的 Luna 版本定价为每百万输入 Token 0.20 美元，每百万输出 Token 1.20 美元。
Given the breakneck speed with which Google has been releasing new Gemini models, it absolutely could have gotten 3.5 Pro out the door at any point this summer. The fact that we have now seen multiple Flash models edging closer to 4.0 suggests Google doesn’t want 3.5 Pro to be compared to the latest releases from OpenAI and Anthropic. Reports suggest that Gemini’s coding capabilities haven’t kept up with recent advances from other AI labs, which comes as Google is also seeing an exodus of AI talent.
考虑到谷歌发布新 Gemini 模型的高速，它完全可以在今年夏天的任何时候推出 3.5 Pro。事实上，我们现在看到多个 Flash 模型不断向 4.0 版本靠拢，这表明谷歌并不希望将 3.5 Pro 与 OpenAI 和 Anthropic 的最新产品进行直接对比。有报道称，Gemini 的编程能力未能跟上其他 AI 实验室的最新进展，而与此同时，谷歌也正面临 AI 人才流失的问题。
Interested parties can begin using Gemini 3.7 Flash today—well, maybe. It’s live in the Gemini API, AI Studio, and Gemini Enterprise. For individuals, the availability is weirdly narrow. Gemini 3.7 Flash is now powering the Gemini Spark agent in the Gemini app, but only if you have an AI Pro or Ultra subscription. It’s not an option in the regular chatbot interface, which continues to run on 3.6 Flash for now.
感兴趣的用户从今天起就可以开始使用 Gemini 3.7 Flash 了——嗯，或许吧。它目前已在 Gemini API、AI Studio 和 Gemini Enterprise 中上线。对于个人用户而言，其可用范围非常有限。Gemini 3.7 Flash 目前为 Gemini 应用中的 Gemini Spark 智能体提供支持，但前提是你必须拥有 AI Pro 或 Ultra 订阅。它在常规聊天机器人界面中尚不可用，该界面目前仍运行在 3.6 Flash 上。