Google announced Gemini 3.7 Flash on August 13, only about three weeks after its prior Flash drop—another sign the company is shipping model updates on a near-sprint schedule. Google says 3.7 Flash brings substantial gains in software engineering, web development, and knowledge work after developer feedback and algorithm work it plans to reuse in future models. Reported benchmark jumps include DeepSWE v1.1 moving from 49.0% to 65.3% and FrontierCode 1.1 Main from 34.4% to 43.6%. On Arena.ai’s WebDev Arena, Elo rose from 1538 (prior Flash) to 1588. Document-heavy work improved on GDP.pdf (34.0% vs 22.0%), and AutomationBench business workflows moved from 17.0% to 30.4%. Google also highlights better multi-step planning, tool use, instruction following, and updated safeguards for CBRN and cyber-offense misuse. Until the end of 2026, introductory API pricing is listed at $0.75 per million input tokens and $3.75 per million output tokens—about half the previous model’s launch price. In the Gemini app, 3.7 Flash is rolling out in Spark for AI Pro and Ultra subscribers first.
Source: https://9to5google.com/2026/08/13/gemini-3-7-flash-launch/