GPT Images 2.5 Prompts and Output Examples

Hermes Agent now supports GPT Image 2.5, letting users edit images with 16 reference images across the Flare and Sunburst versions.

The author uses the MoodLock plugin to convert an AI-generated cosplay outfit into a reusable garment prompt and shares a full ChatGPT Images 2.5 prompt example.

The author says GPT Image 2.5 can upgrade old video thumbnails to a new version in one click, and calls the result extremely impressive.

A user posted comparison images testing how "intelligent" GPT Image 2.5 is and concluded the results were not great.

An Arabic-language author shared a face-to-face video comparison showing whether GPT-Image 2.5 keeps the same character's facial features consistent across images, unlike older models.

Luma says its Agents now run two new OpenAI image models: Flare for speed and volume, and Sunburst for precise edits where you bring a reference, change what's off, keep the rest, then move it into video.

The author posts an image generated by GPT Image 2.5 as a stress test, asking whether it got every detail right and inviting readers to judge for themselves.

An editor with 20 years at women's fashion magazines says AI images never beat real fashion photography, but GPT Image 2.5 is the first to feel like it might step onto the same ring.

A Japanese user shares LINE stickers made with Image 2.5 and includes the reusable prompt.

The author used GPT Image 2.5 on the third-party platform Blyth to generate a game scene concept image and shared a link to try it.

The author calls Images 2.5 the best image model right now but says it still lacks click-an-object-to-edit-it interaction, and asks OpenAI to add it.

The poster shares a batch of sample images generated with GPT Image 2.5 and says a full review is coming that evening.