GPT Image 2.5 Image Generation — Flare and Sunburst, five refinement tiers, transparent output
OpenAI's GPT Image 2.5 comes in two variants on XinYu AI — Flare and Sunburst. Five refinement tiers from Low to Max, up to 16 reference images per run, and PNG output with a real alpha channel. Both text-to-image and reference-based editing are supported, on the same credit balance as every other model.
Available builds
- GPT Image 2.5 FlareThe faster variant and the everyday pick; five refinement tiers, 1K/2K/4K, up to 16 reference images, transparent output
- GPT Image 2.5 SunburstSame image quality as Flare at about twice the time, tuned for edit precision; identical parameters and pricing
What it's good at
Five refinement tiers, not three
Low / Medium / High / Ultra / Max. The tier decides how much compute goes into the image, and the panel shows the exact credit cost before you run it.
Transparent output, natively
Pick Transparent and you get a PNG with a real alpha channel — everything outside the subject is genuinely transparent, not white. That removes a cutout step for product shots, stickers and asset pieces.
Up to 16 reference images
More room for character consistency, multi-image blends and style references — and both variants edit with references.
Why run it on XinYu AI
One balance, every model
Image, video, audio and text draw from the same credit balance — no separate subscription per model. Credits never expire.
Chain it on one canvas
A generated image becomes the next step's reference; a clip can be trimmed, extended and scored right there — no export/import round trips.
Drive it from MCP or the CLI
Generate straight from Claude or Cursor, or wire the CLI into your own scripts and automation. The browser is optional.
FAQ
Flare or Sunburst?
Same image quality, same parameters, same price — the difference is speed and positioning. Flare renders about twice as fast and is the everyday pick; Sunburst is slower and tuned for edit precision, which suits retouching and fine detail.
Are the five tiers the same as GPT Image 2's three?
No. Each generation splits its tiers by its own compute, so "High" on 2.5 does not correspond to "High" on GPT Image 2. For the best 2.5 output, pick Max.
How do I get a transparent background?
Set Background to Transparent in the image node's parameter panel and the result is a PNG with a real alpha channel. The control appears only on GPT Image 2.5 — no other image model here has it.
Do 1K / 2K / 4K support the same aspect ratios?
Almost identical. 1:1, 4:3, 3:4, 3:2, 2:3, 16:9, 9:16 and 21:9 are available at 1K, 2K and 4K; 2K additionally offers 9:21. The ratio you choose is the ratio you get.
Are reference images charged extra?
Yes. The upstream bills per reference image, so attaching references counts toward the price. The number in the panel tracks the reference count live, so you see the final amount before you run it. See the pricing page for tier prices.
Try it now
New accounts start with free credits — no card up front. Image, video and audio share one balance, and credits never expire.