← Back to all models
GPT Image 2.5
OpenAIImage GenerationProprietaryOpenAI's current image generation line, released September 2026 in two variants: Flare (fast default, ~50% lower latency than GPT Image 2) and Sunburst (precision editing). Six quality tiers from $0.006 to $0.211 per 1024x1024 image.
Abilities
Image GenerationMultilingual
Use Cases
CreativeContent WritingEnterprise
Available in Tools
ChatGPThttps://chat.openai.com
Availability
ChatGPT Webhttps://chat.openai.com
How to Use
Use the OpenAI Images API with model ID `gpt-image-2.5-flare` (fast default) or `gpt-image-2.5-sunburst` (slower, precision editing). Pick one of six quality tiers (low/medium/high/xhigh/max/auto) to trade cost against fidelity.
Pros
- Two specialized variants — Flare for speed, Sunburst for precision editing
- ~50% lower latency than GPT Image 2 on the Flare variant
- Six quality tiers — from $0.006 per image for drafts to $0.211 for maximum fidelity
- Native successor to DALL-E 3 and GPT Image 1/2 with image input and editing support
Cons
- Closed source — no weights, LoRAs, or self-hosting
- Sunburst variant is noticeably slower than Flare
- Top quality tiers get expensive at scale ($0.211 per image at max)
What to Use It For
Perfect For
Fast conversational image generation
Flare halves latency vs GPT Image 2 while keeping ChatGPT integration
Precision image editing
Sunburst is purpose-built for controlled, localized edits
Good For
Cost-tiered image pipelines
Six quality tiers let you pay for fidelity only where it matters
Not Recommended
Custom style fine-tuning
No LoRAs or fine-tuning — open-weight models offer full style control
Try instead: FLUX.1, Stable Diffusion 3.5
Do Not Use For
Video generation
Still-image model only — OpenAI has no video model since Sora shut down
Try instead: Veo, Kling 2.5 Turbo Pro
Technical Details
Pricing$0.006 – $0.211 per 1024x1024 image (text input $5/1M, image input $8/1M, image output $30/1M tokens)