Google Ships Gemini 3.7 Flash, Its Cheapest Workhorse Model for Coding and Agents

Google unveiled Gemini 3.7 Flash on August 13, pitching it as its most intelligent workhorse model yet for coding and agentic workflows, just three weeks after Gemini 3.6 Flash. The model arrives at an introductory price of $0.75 per million input tokens and $3.75 per million output tokens — half the launch price of its predecessor — available through the end of 2026 before reverting to $1.50 and $7.50 on January 1, 2027, the company said in a blog post covered by Decrypt, The Verge and Search Engine Journal.

A Cheaper Brain for Autonomous Agents

Gemini 3.7 Flash accepts up to one million input tokens, roughly 750,000 words, and returns 64,000 output tokens, handling text, images, video, audio and PDFs while supporting tool calls and computer control. Google says the model leads on 11 of 13 internal benchmark comparisons against GPT-5.6 and rivals, including top scores in web development and AutomationBench for enterprise workflows. It is available immediately in more than 160 countries through Google Antigravity, the Gemini API, Google AI Studio and Android Studio, and now powers Spark, the 24/7 personal agent for Google AI Pro and Ultra subscribers.

Rolling Out Fast

Within a day, the model became selectable in AI Mode for Search for AI Pro and Ultra subscribers, with Google Search product VP Robby Stein saying it is better at following instructions and understanding intent. The release landed the same day OpenAI previewed GPT-5.6 Sol Ultrafast, a Cerebras-powered tier that runs its top model up to 14 times faster — but Google's model is generally available while OpenAI's is invite-only, underscoring a speed race that has shifted from raw intelligence to real-time agents.