Gemini API: gemini-3.1-flash-lite-preview Shut Down — Migration to GA Version Required

What Changed in the Gemini API on May 25, 2026?

Google shut down the preview version of the gemini-3.1-flash-lite-preview model in the Gemini API on May 25, 2026. According to the Gemini API Changelog, development teams must use the stable GA (Generally Available) version gemini-3.1-flash-lite instead of the preview endpoint, which is no longer available.

The transition period was relatively short — just 18 days between the GA release and the shutdown of the preview.

What Is the Timeline of This Shutdown?

The Gemini API Changelog documents the following sequence of events:

• May 7, 2026 — Google releases the GA version gemini-3.1-flash-lite as a stable, production-ready model optimized for speed, scalability, and cost
• May 7, 2026 — A deprecation notice for the preview version is simultaneously announced
• May 25, 2026 — The preview endpoint gemini-3.1-flash-lite-preview is officially shut down

How to Perform the Migration and Why Is the GA Version the Better Option?

Migration is technically straightforward: in all API calls, simply replace the model ID from gemini-3.1-flash-lite-preview to gemini-3.1-flash-lite. The GA version uses the same parameters, the same API conventions, and the same capabilities as the preview, but with additional stability and availability guarantees that Google provides for production models.

gemini-3.1-flash-lite is designed for applications requiring high throughput at lower costs — positioned below gemini-3.1-flash in Google’s model hierarchy, but offering a more favorable performance-to-cost ratio for simpler tasks.

Development teams that used the preview for testing should now check all production integrations and CI/CD pipelines to ensure no old model IDs remain that could cause errors.