Lovable Adds Gemini 3.6 Flash, Makes It the Default Model

LovableView original changelog

Lovable added Google's Gemini 3.6 Flash as a model option for apps' built-in AI features, supporting text, image, audio, and video input for fast coding, reasoning, and agentic workflows. One day later, Lovable made it the default model for any app that calls built-in AI features without specifying one, replacing the earlier Gemini 3 Flash Preview default.

Key Takeaways

  • Gemini 3.6 Flash is now selectable for Lovable apps' built-in AI features, supporting text, image, audio, and video input.
  • One day after launch, it became the default model for AI Connector calls that don't specify one, replacing Gemini 3 Flash Preview.
  • Apps that already pin a specific model are unaffected by the default change; the upgrade is automatic for everyone else.
  • The model targets the AI Connector, a runtime service end users interact with, distinct from the agent that helps build the app during development.
  • Gemini 3.6 Flash joins three GPT-5.6 variants (Sol, Terra, Luna) as part of Lovable's expanded in-app model lineup.
  • The rollout landed within days of GPT-5.6's own release inside Lovable, reflecting a fast-follow pattern on frontier model access.

Sources & Mentions

1 external resource covering this update


A New Flash Model, Then the New Default

Lovable rolled out Gemini 3.6 Flash, Google's latest-generation Flash model, as an option for the AI Connector, the runtime service that powers AI features inside apps built on Lovable. It accepts text, images, audio, and video as input and responds with text, positioned for fast coding, reasoning, and agentic workflows. The next day, Lovable made it the default: apps that call built-in AI features without specifying a model now use Gemini 3.6 Flash automatically, replacing the previous Gemini 3 Flash Preview default.

What Changed for Builders

Apps that already specify a model explicitly are unaffected by the default change. For everyone else, the switch is automatic and required no action. Gemini 3.6 Flash joins the three GPT-5.6 variants (Sol, Terra, Luna) as model options builders can choose to power an app's AI features.

Why It Matters

This is distinct from the model that powers Lovable's own build agent; it's about the AI Connector that end users' apps call at runtime. Making Gemini 3.6 Flash the new default means every app relying on the default model gets a multimodal capability and speed upgrade automatically, without the builder having to opt into a specific release.


Mentioned onAlphaSignal
Lovable Adds Gemini 3.6 Flash as Default AI Model | Yet Another Changelog