Google previews Gemini 3.1 Flash-Lite on Vertex AI

Google made Gemini 3.1 Flash-Lite (gemini-3.1-flash-lite) available in public preview on March 03, 2026, describing it as its most cost-efficient Gemini model. The model is optimized for low latency use cases involving high-volume, cost-sensitive LLM traffic.

The same update also deprecated several video generation preview endpoints and listed recommended replacements, with a migration deadline before April 2, 2026. Separately, Google said Vertex AI documentation is no longer being updated because Vertex AI services are now part of Gemini Enterprise Agent Platform.

Recent Vertex AI release notes show a broader shift toward Gemini and Agent Platform capabilities, including Gemini 3.1 Pro preview in Model Garden, Gemini 3.1 Flash Image in public preview, updates to Workbench, and new agent, RAG, evaluation, and Model Garden features.