MDL-4632EST.2026 · IDX.068
Language modelIn production

Gemini 3.7 Flash

Google DeepMind · USA / UK · 2026

Google's fastest coding-and-agents model, sold at half price until the end of 2026.

wujec.ai score

/10

Community score

no votes yet
Sign in to rate

Gemini 3.7 Flash, announced on 13 August 2026, is Google's newest Flash-class model and the first the company positions primarily around coding and agent work rather than raw speed. Against Gemini 3.6 Flash it lifts FrontierCode 1.1 from 34.4% to 43.6%, DeepSWE v1.1 from 49.0% to 65.3% and AutomationBench from 17.0% to 30.4%, while WebDev Arena rises from 1538 to 1588 Elo. It keeps the Flash-line million-token context window, accepts text, images, video, audio and PDFs, and returns text only. The pricing deserves a footnote. Google advertises USD 0.75 per million input tokens and USD 3.75 per million output as "half" the original 3.6 Flash rate, but that is an introductory rate written into the price list as expiring on 31 December 2026, after which both models move to USD 1.50 and USD 7.50. Gemini 3.6 Flash currently carries exactly the same discounted rate, so on 15 August 2026 the two models cost the same to run. The model is available in Google Antigravity, the Gemini API through AI Studio and Android Studio, the Gemini Enterprise Agent Platform, and in Gemini Spark for AI Pro and Ultra subscribers in more than 160 countries.

#fast#coding#agentic#multimodal#efficient
Official website

News

Videos

No videos yet.

Reviews

No reviews yet. Be the first!

Sign in to write a review