Google released Gemini 3.7 Flash on August 13, 2026—just three weeks after Gemini 3.6 Flash—as an upgrade optimized for coding, agentic tool use, and enterprise document workflows. The model launches at introductory pricing of $0.75/1M input tokens and $3.75/1M output tokens through December 31, 2026, then doubles to $1.50/$7.50 on January 1, 2027.
Coding gains are substantial. Gemini 3.7 Flash scores 65.3% on DeepSWE v1.1 (up from 49.0%), 43.6% on FrontierCode 1.1 Main (up from 34.4%), and 30.4% on AutomationBench (up from 17.0%). On the Artificial Analysis Intelligence Index, it scores 56—just behind frontier models like GPT-5.6 Terra (57) and Muse Spark 1.2 (57). It did not top every benchmark: Claude Sonnet 5 leads on some agentic tasks, and OSWorld 2.0 remains GPT-5.6 Terra's domain.
Speed and cost-per-task are the standout moves. At 340 output tokens per second, Gemini 3.7 Flash is nearly 3x faster than GPT-5.6 Terra and places it on the Intelligence vs. Time per Task Pareto frontier with an average task time of 1.7 minutes (40% faster than Terra at max reasoning). Per-task cost drops 30% versus 3.6 Flash despite identical token pricing, reflecting fewer turns and better efficiency.
For developers, the timing matters: the introductory rate expires at year-end 2026. After January 1, 2027, input and output pricing double. Even at the higher rate, Gemini 3.7 Flash remains cheaper than Claude Sonnet 5 and GPT-5.6 Terra. This rapid release cycle—three Flash tiers in three months—reflects Google's competitive positioning: pushing frequent model updates rather than waiting for major generational jumps, targeting teams building coding agents and multi-step automation before the market consolidates around a single default.