Google rolls out new Gemini Flash models and announces 'Frozen v2' AI chip
Alphabet’s Google announced a trio of cheaper Gemini models on July 21. The lineup includes Gemini 3.6 Flash, Gemini 3.5 Flash‑Lite and Gemini 3.5 Flash‑Cyber. The Flash series is positioned to cut token usage (up to 65% fewer tokens on some tasks), lower latency and reduce costs for large‑scale AI agents. Gemini 3.5 Flash‑Cyber is a cybersecurity‑focused model that runs with the CodeMender agent and is being piloted for governments and trusted partners.
The company also confirmed that the flagship Gemini 3.5 Pro, originally slated for a June launch, has been delayed because the model fell short of internal targets, particularly in code‑generation performance. Google’s CEO Sundar Pichai highlighted potential annual savings of up to $1 billion for customers shifting workloads to the Gemini family.
In parallel, Google disclosed development of an internal AI server chip dubbed “Frozen v2.” Engineers say the chip could deliver six to ten times more tokens per unit of power than current TPUs, aiming to ease a chronic AI compute shortage that has forced Google Cloud to turn away customers. The chip, designed to run Gemini models directly in silicon, is planned for rollout around 2028 and contributed to a modest rise in Alphabet’s share price.