Google发布了Gemini 3.7 Flash大语言模型，这是其Flash系列的最新版本。[1]该模型专为代码编写和AI Agent应用优化，相比三周前推出的3.6 Flash版本在多个关键领域取得了显著性能提升。[1]
在代码相关任务上，Gemini 3.7 Flash表现突出。[1]FrontierCode 1.1 Main基准测试中该模型得分为43.6%，较3.6 Flash的34.4%提高了9.2个百分点。[1]DeepSWE v1.1基准测试结果为65.3%，相比3.6 Flash的49.0%有明显提升。[1]在知识密集型任务上，GDP.pdf基准测试中新版本达到34.0%，而旧版本为22.0%。[1]自动化方面，AutomationBench测试中3.7 Flash得分30.4%，对比3.6 Flash的17.0%；网页开发方面，WebDev Arena Elo评分则为1588，相比3.6 Flash的1538。[1]
在定价策略上，Google采取了更具竞争力的方案。[1]该模型的推介价格为每百万输入token 0.75美元、每百万输出token 3.75美元，定价有效期至年底。[1]新模型将通过Google Antigravity、Gemini API、Google AI Studio、Android Studio及Gemini Enterprise等平台提供。[1]
Google has launched Gemini 3.7 Flash, the latest addition to its Flash series of AI models, arriving three weeks after the release of Gemini 3.6 Flash [1]. The new model is optimized for code writing and AI agent applications, delivering substantial performance improvements across multiple benchmarks [1].
On the FrontierCode 1.1 Main benchmark, Gemini 3.7 Flash achieved 43.6% compared to 34.4% for its predecessor [1]. The model also showed gains in code-related tasks, scoring 65.3% on DeepSWE v1.1 versus 49.0% for version 3.6 [1]. Additional improvements include a WebDev Arena Elo rating of 1588 against 1538, GDP.pdf benchmark performance of 34.0% compared to 22.0%, and AutomationBench results of 30.4% versus 17.0% [1].
Google has priced Gemini 3.7 Flash at $0.75 per million input tokens and $3.75 per million output tokens as an introductory rate through the end of the year [1]. The model is available through Google Antigravity, the Gemini API, Google AI Studio, Android Studio, and the Gemini Enterprise platform [1].