GPT Image 2.5 Flare Edit

GPT Image 2.5 Flare Edit makes precise, instruction-scoped edits to an existing image while the subject, the composition and the background stay intact.

By OpenAI

Pricing: $0.0611 per image

Overview

GPT Image 2.5 Flare Edit applies precise edits to an existing image from a plain-language instruction. It changes only what the instruction asks for and holds the subject, the composition and the background steady. Reference subjects stay recognizable across styles and across successive edits.

Key capabilities

• **Six quality tiers**, pick low through max to trade detail against cost per image.
• **Multi-region failover**, Runflow runs inference across several regions. If one region has an outage, traffic routes to another automatically.
• **Multi-image reference**, pass up to 16 reference images for composition-aware editing.
• **Optional mask**, target one region of the image with `mask_url` for precise inpainting.

GPT Image family

| Model | Details |

|---|---|

| **GPT Image 2.5 Flare Edit** (this model) | It changes only what the instruction asks for and holds the subject, the composition and the background steady. Reference subjects stay recognizable across styles and across successive edits. |

| **GPT Image 2.5 Sunburst Edit** | Sibling editor tuned for tightly scoped instructions |

| **GPT Image 2 Edit** | Previous generation editor |

Technical specifications

| Spec | Details |

|---|---|

| Quality tiers | low, medium, high (default), xhigh, max, auto |

| Output formats | png, jpeg, webp |

| Inputs | prompt, image_urls (required), optional mask_url |

| Reference images | up to 16 per request |

| Images per request | 1 to 4 |

| Pricing | $0.0143 low, $0.0216 medium, $0.0611 high, $0.1021 xhigh, $0.2191 max, per image |

Examples

• Replace the background with a sunlit seaside cafe at golden hour, keep the subject's pose, outfit, and lighting on the face intact

Frequently asked questions

What does GPT Image 2.5 Flare Edit do?

GPT Image 2.5 Flare Edit edits an existing image from a plain-language instruction. Send the image in image_urls and the instruction in prompt. The model applies the change and returns a new image.

How much does one image cost?

Price follows the quality parameter: $0.0143 per image at low, $0.0216 at medium, $0.0611 at high, $0.1021 at xhigh and $0.2191 at max. High is the default.

How many reference images can I send?

Up to 16 URLs in image_urls. The model reads all of them for composition-aware editing, so extra references help when the edit has to keep a subject recognizable.

Can I edit only part of the image?

Yes. Pass a mask image in mask_url. White areas mark the region to edit and the rest of the image stays as it is.

How many images can one request return?

Between 1 and 4. Set num_images. Each returned image bills at the rate for the quality tier you selected.

What output formats are available?

png (default), jpeg and webp. For jpeg and webp you can also set output_compression between 0 and 100.

Related models

• GPT Image 2.5 Sunburst Edit, GPT Image 2.5 Sunburst Edit scopes every edit tightly to the instruction and holds the subject and the composition steady across many rounds of revision.
• FLUX.1 [dev] Edit, Image-to-image edits with FLUX.1 [dev]. Pass a source image plus a prompt and a transformation strength to redirect the composition while preserving subject identity.
• Multi Angles, Re-render the same subject from a different camera angle. Pick a horizontal view, a vertical angle, and a shot distance: the model rotates the camera around the subject and re-renders the scene from that pose.
• FLUX.2 [klein] 4B Edit, Fast image-edit endpoint backed by FLUX.2 [klein] 4B. Pass a base64 source image with the prompt for low-latency edits.

Discoverable surfaces

• Dispatch endpoint: POST https://api.runflow.io/v1/models/gpt-image-2.5/flare/edit/runs
• Per-model spec (markdown): https://app.runflow.io/models/gpt-image-2.5/flare/edit/llms.txt
• Docs page: https://docs.runflow.io/models/gpt-image-2.5/flare/edit
• Public OpenAPI spec: https://docs.runflow.io/api/openapi.public.json
• Agent skill (start here): https://www.runflow.io/.well-known/agent-skills/runflow/SKILL.md