OpenAI's image generation and editing model in ChatGPT and the API, with Sunburst and Flare variants.
Updated EU status checked Last shown
Sources: creator videos at the linked timestamps. EU status is information, not legal advice; see how we check.
ChatGPT Images 2.5 is OpenAI's image generation and editing model, announced on 8 September 2026 as the successor to Images 2.0. OpenAI describes improvements in detail, lighting and texture, preservation of subjects from reference photos, and more reliable instruction-following across multi-turn edits. It cites up to 50% lower generation latency than Images 2.0.
In ChatGPT it adds Sketch (draw a reference via "@Sketch"), templates for formats such as posters and merch, comments placed directly on images for targeted edits, and shareable prompts. The model also handles more complex layouts, including transparent backgrounds. Creators reviewing it report only slight gains over GPT Image 2 in many tests, with remaining failures on some text-heavy or diagram-style prompts such as signs, chess diagrams and biology labels.
For developers there are two API models: GPT-Image-2.5 Flare (faster, same quality and editing improvements) and GPT-Image-2.5 Sunburst (extra precision for detailed work, longer generation times). ChatGPT access is offered to all ChatGPT, ChatGPT Work and Codex users on desktop, mobile and web. Creator-reported API pricing ranges from roughly $0.005 to $0.21 per 1024 image depending on variant and settings; check OpenAI's pricing page for current figures.
Watch the moment: the exact second top creators show it. Videos load only when you press play, via youtube-nocookie.com.
New OpenAI image model in ChatGPT with improved consistency when editing from reference photos, plus a new Sketch feature to draw a reference. Creator demoed sketch-to-realistic. Available across tiers and in the API in two variants.
Creator tests OpenAI's new image model on many hard prompts against GPT Image 2 and Nano Banana 2. New features include sketch-to-photo, annotation editing, consistent multi-turn editing and transparent layers exported to PSD. It is slightly better than GPT Image 2 in most tests but still fails on gibberish signs, chess diagrams, frogs and biology labels. Two variants: Flare (faster, lower quality) and Sunburst (premium). It is ranked #1 on text-to-image and editing arenas, with API pricing from about $0.005 to $0.21 per 1024 image.
Closest neighbours by meaning, not just by category.
Model · OpenAI
OpenAI's image generation model, accessed via the API and ChatGPT, supporting text prompts and reference images.
Model · OpenAI
OpenAI's image generation and editing model family, accessed via the API and ChatGPT.
Model · OpenAI
OpenAI image generation and editing model with GIF output, SVG export and image markup, in Flare and Sunburst variants.
App · OpenAI
OpenAI's chat assistant for answering questions, writing, image generation, coding and agent-style tasks, on web and apps.
Model · Microsoft
Microsoft AI's text-to-image and image-editing model, with a lower-cost Flash variant, offered via Microsoft Foundry.
Model · Microsoft
Microsoft AI's text-to-image and image-editing model in the MAI Image family, with accurate text rendering.
