Google employees are already testing the next Gemini Flash AI model
Executive Take
Google is competing on release speed and cost, not raw capability, betting that cheaper agentic tools win more customers than the smartest model. Buyers should stop assuming "best model" and start comparing cost per task.
Executive Summary
Google employees are internally testing "Gemini 3.8 Flash Preview" on the company's Jetski coding platform, just weeks after releasing Gemini 3.7 Flash. Google lacks a frontier model as OpenAI and Anthropic lead, and is instead pushing fast, cheap Flash models roughly monthly, per CEO Sundar Pichai.
Why It Matters
AI and technology leaders choosing model vendors need to track this shift toward cheap, fast models for coding and agents, since token cost now drives enterprise AI budgets as much as raw quality does.
Bizquad Perspective
Google's monthly Flash cadence signals it has conceded the frontier race and is instead trying to make itself the default cheap infrastructure layer for enterprise AI agents.