Back to rankings

Lovable: GPT Image 2.6 Flare and Sunburst for App AI Features

LovableView original changelog

Lovable added two OpenAI image models, GPT Image 2.6 Flare and GPT Image 2.6 Sunburst, as options for the AI features built into apps made with Lovable. Flare is the faster model for previews and in-app image editors, while Sunburst targets detailed visuals and precise edits. Both can make short animated GIFs, guide edits with a drawing on the image, and output transparent PNG or WebP stills. They are not available while EU inference is enabled.

Key Takeaways

  • Two image models from OpenAI, GPT Image 2.6 Flare and Sunburst, are now available for the AI features inside apps built with Lovable.
  • Flare is described as the faster model, suited to previews and in-app image editors, while Sunburst targets detailed visuals and precise edits.
  • Both models can produce short animated GIFs, extending app image features beyond still pictures.
  • Edits can be guided by a drawing made on top of the image.
  • Both models can output a still image with a transparent background, as PNG or WebP.
  • Image generation and editing use credits, and neither model is available while EU inference is enabled.

Two New OpenAI Image Models for App AI Features

Lovable added two OpenAI image models to the AI features that apps built with Lovable can include: GPT Image 2.6 Flare (openai/gpt-image-2.6-flare) and GPT Image 2.6 Sunburst (openai/gpt-image-2.6-sunburst). Builders can ask Lovable to use either one when it builds an image feature, so an app can generate and edit images with these models.

Flare and Sunburst

The changelog describes Flare as the faster model, suited to previews and in-app image editors. Sunburst is aimed at detailed visuals and precise edits.

Shared Capabilities

Both models support the same set of image features:

  • Producing a short animated GIF.
  • Guiding an edit with a drawing made on top of the image.
  • Outputting a still image with a transparent background, as PNG or WebP.

Credits and Availability

Generating or editing images with either model uses credits. Neither model is available while EU inference is enabled, so apps that have turned on EU inference cannot use them. Lovable's documentation covers both topics on its AI features page and its EU inference page.