Overview
- Google published Gemini 3.7 Flash on Thursday, August 13, 2026, calling it a faster, cheaper “workhorse” model tuned for coding, multi‑step agent tasks and complex document understanding.
- Official benchmark results published by Google show large year‑over‑year jumps versus 3.6 Flash, including FrontierCode rising to 43.6% from 34.4% and DeepSWE to 65.3% from 49.0%, plus gains on WebDev Arena and AutomationBench.
- Google set an introductory API price through December 31, 2026 of $0.75 per 1M input tokens and $3.75 per 1M output tokens and multiple reports say those rates are expected to double after the promotion ends on January 1, 2027.
- The model is rolling out across Google platforms today — Gemini Spark for Pro/Ultra subscribers, the Gemini API, Google AI Studio, Android Studio, Antigravity and enterprise tools — and supports text, image, video, audio and PDF inputs, function calling, large token contexts and configurable 'thinking_level' settings.
- The rapid three‑week cadence of Flash updates reflects a product strategy to ship practical, low‑latency infrastructure while the premium Gemini 3.5 Pro remains unreleased and Google adjusts strategy after recent DeepMind leadership changes and competitive moves from other AI labs.