Google releases Gemini 3.7 Flash, its cheapest coding model

Google released Gemini 3.7 Flash on 13 August 2026, three weeks after Gemini 3.6 Flash. It calls the model 'our most intelligent workhorse model yet for coding and agents.'

Introductory pricing is $0.75 per million input tokens and $3.75 per million output, through 31 December 2026, then $1.50/$7.50 from 1 January 2027 — the same rate Google then applied to Gemini 3.6 Flash. On Google's benchmarks: FrontierCode 1.1 Main rose from 34.4% to 43.6%, DeepSWE v1.1 from 49.0% to 65.3%, GDP.pdf from 22.0% to 34.0%, and AutomationBench from 17.0% to 30.4%.

Gemini 3.5 Pro, Google's flagship announced in May 2026, remains unreleased after missing three internal deadlines; a July report attributed one miss to hallucinations and called the Flash line stopgap releases. On 5 August 2026, DeepMind CEO Demis Hassabis stepped back to chairman, handing Gemini development to CTO Koray Kavukcuoglu; Jeff Dean left the same day to found Discovery Loop.

xAI shipped Grok 4.6 the day before, on 12 August, at $2/$6 per million tokens with an Artificial Analysis Intelligence Index score of 61 against Gemini 3.7 Flash's 56. Google's August update also included the Pixel 11 phones, new Gemini Live features and a free year of Google AI Pro or Plus for eligible students from 19 August.

Alongside the model, Google shipped Gemini Omni 1.1 Flash for video: scene extension, first-and-last-frame interpolation and 4K upscaling. Gemini Live also gained Personal Intelligence, a Daily Brief, the Spark agent and hands-free inbox management. Eligible students could claim a free year of Google AI Pro (US) or Google AI Plus (140+ countries) from 19 August 2026, though billing starts automatically a year after signup.