ChatGPT Images 2.5 Image Generation & Editing Model
OpenAI released ChatGPT Images 2.5 on 8 September 2026 as the successor to Images 2.0: sharper detail, more natural lighting and texture, better preservation of the people and products in your reference photos, edits that stay consistent across long multi-turn sessions, up to 50% lower generation latency, and two new API models in GPT-Image-2.5 Flare and Sunburst
Tool Interface
Interactive tool will be available soon
Features
- ✓ Sharper detail and more natural lighting and texture, with generation latency reduced by up to 50% versus Images 2.0 per OpenAI
- ✓ Better at preserving the people and products in your reference photos, so personal and on-brand assets need less retouching
- ✓ More reliable editing across long conversations: earlier changes are more likely to survive later edits and quality holds up over time
- ✓ Comment-based edits let you target one spot instead of redrawing the whole image, alongside Sketch drawing inside ChatGPT and templates for common formats
- ✓ Two API models: Flare for the same quality and speed gains, Sunburst for higher precision on detailed creative work at longer generation times; prompts can be shared for others to reproduce
How to Use
- Read the official launch page at https://openai.com/index/introducing-chatgpt-images-2-5 for capabilities, rollout and pricing
- In ChatGPT, ChatGPT Work or Codex, drop in a reference photo or draw a Sketch, then describe the image you want
- Iterate with comment-based edits that target only what should change, so earlier details are not lost to a full redraw
- Developers pick gpt-image-2.5 flare or sunburst in the API and keep the model name in configuration for easy switching and rollback
FAQ
What is ChatGPT Images 2.5?
It is OpenAI's image generation and editing model released on 8 September 2026; the official page is https://openai.com/index/introducing-chatgpt-images-2-5 . OpenAI says it improves detail, lighting and texture, fidelity to subjects in reference photos and consistency across multiple edits, while cutting generation latency by up to 50% versus Images 2.0. It rolled out the same day to ChatGPT, ChatGPT Work and Codex across all tiers.
What changed most versus Images 2.0?
OpenAI lists sharper detail, more natural lighting and richer textures, better preservation of the people and products in your reference photos, more reliable editing across multiple turns, and up to 50% lower image generation latency. In ChatGPT it also adds Sketch drawing, comment-based edits and templates for popular formats.
How do comment-based edits work?
Per OpenAI, you place comments directly on an image to request a change at a specific spot, and the model edits only what you pointed at rather than rebuilding the whole asset. OpenAI also says Images 2.5 follows specific editing instructions more reliably across longer conversations, so earlier changes stay consistent.
Should I use Flare or Sunburst?
Both are GPT-Image-2.5 models in the API. OpenAI describes Flare as bringing the same quality, editing and speed improvements as the ChatGPT experience, while Sunburst adds precision for detailed creative work with longer generation times. Choose Flare for throughput, Sunburst when the final asset must be precise.
Is generation actually faster?
OpenAI states latency is reduced by up to 50% compared with Images 2.0 so you can generate and refine concepts more quickly. Real-world speed varies with prompt complexity, the number of reference images and service load, so check the official release notes and your own benchmarks.
Can I use the images commercially?
OpenAI's official terms and content policies are the source of truth: commercial use and content rules differ by plan and region, and enterprise workspaces may depend on an admin setting. Always confirm against the latest OpenAI terms of service and content policy before publishing.