Particle.news

Google Releases Cheaper Gemini Flash Models as Flagship Pro Remains Delayed

The new low-cost, domain-targeted Flash variants are meant to cut enterprise AI spending because the Gemini 3.5 Pro is still in partner and government testing.

Overview

  • Google announced three lighter Gemini variants on Tuesday that add Gemini 3.6 Flash, Gemini 3.5 Flash-Lite, and a cybersecurity-focused Gemini 3.5 Flash Cyber to its product line.
  • The company said Gemini 3.5 Pro is still being tested with partners and regulators and gave no firm launch date after the model missed internal targets, especially on coding performance.
  • Developers spotted internal model IDs such as gemini-3.6-flash-tiered in Google’s Antigravity IDE, which suggests tiered access and that the models may soon appear in AI Studio and the Vertex API.
  • Google and DeepMind pitched Flash Cyber as a lower-cost tool for finding and fixing security flaws, and Google highlighted Flash pricing and reduced token use as a way to lower running costs for customers.
  • The move gives Google a short-term commercial play against rival releases and regulatory scrutiny by offering cheaper, task-specific models while it refines higher-capability frontier models.