# Models updated: Gemini 3.8 Flash and GPT-5.6 Luna

Published: 2026-09-28 | Source: https://geoready.co/news/flash-tier-everyday-defaults

The flash tier now runs the models people get by default today: Gemini 3.8 Flash, which the Gemini app and Google's AI Mode run, and GPT-5.6 Luna, ChatGPT's default on the Free and Go plans.

GEO is about scoring your brand on the answers real users get. So the matrix follows the models the apps give people by default, and we update it when they change.

## What changed

- **Gemini 3.8 Flash** replaces Gemini 3.5 Flash. Google released it on September 2, and it now runs in the Gemini app and in AI Mode in Google Search.
- **GPT-5.6 Luna** replaces GPT-5.5 as OpenAI's flash-tier model. It is what ChatGPT's Free and Go plans run by default.

The top tier (Gemini 3.1 Pro, GPT-5.6 Sol and Claude Sonnet 5), Claude Haiku 4.5 and the Mistral models are unchanged. GPT-6 is rolling out in ChatGPT Work and Codex, and we will move the matrix when it becomes the everyday default in chat.

## What it means for you

Historical audits keep their scores and still show the models they ran on. New audits use the updated matrix automatically. Your score chart marks the date, so a jump after it may come from the new models rather than from changes to your site.