AI Daily Brief — 3 March 2026

Google ships the Gemini 3.1 Flash-Lite preview — pitched as the most cost-effective Gemini model yet, available immediately through the Gemini API in Google AI Studio and Vertex AI.

Top stories

• Gemini 3.1 Flash-Lite preview. The new low-latency tier sits below Flash and Pro in the Gemini line, targeting high-volume classification, extraction and short-form generation use cases. via Google
• Pricing & availability. Preview pricing keeps Flash-Lite as the cheapest token-economics tier in the Gemini stack; general availability planned later in Q2 2026. via Google Cloud

By the numbers

• Gemini 3.1 tier order from highest to lowest cost: Pro · Flash · Flash-Lite

Compiled by AI Feed’s editor from publicly reported announcements on 3 March 2026.