Gemini 3.7 Flash

Gemini 3.7 Flash went generally available on August 13, 2026, three weeks after Gemini 3.6 Flash and with Gemini 3.5 Pro still unshipped. The model id is gemini-3.7-flash. Introductory pricing is $0.75 per million input tokens and $3.75 per million output through December 31, 2026, exactly half the 3.6 Flash rate, after which it reverts to $1.50/$7.50. Batch and Flex processing halve that again to $0.375 and $1.875. It carries a 1,048,576 token input window, up to 65,536 output tokens, and accepts text, image, audio, video, and PDF input. Google says it did not retrain from scratch: 3.7 Flash is built on algorithmic improvements and user feedback applied to the 3.6 line, and it fully replaces rather than supplements the previous version. The published gains are concentrated in coding and agent work. DeepSWE v1.1 moves from 48.6 to 65.3 percent, FrontierCode 1.1 Main from 34.4 to 43.6, Terminal-Bench 2.1 lands at 85.8, and GDM-MRCR v2 reports 97.0 percent long-context retrieval at 128K. The honest caveat is the same one that applied to 3.6: these are Google-reported figures, Artificial Analysis has no SWE-bench Verified score for 3.7 Flash yet, and independent runs have not landed. The strategic read matters as much as the numbers. Google is shipping its cheap tier on a three-week cadence at falling prices while its flagship Pro model slips a third and fourth time, which makes Flash the practical Google model for anyone building agents today. Day-one availability spans the Gemini API, Google AI Studio, Android Studio, Google Antigravity, the Gemini Enterprise Agent Platform, and the Gemini app.