Reeldo AIReeldo AI

OpenAI GPT Image 2.6 · Kie · Reeldo

GPT Image 2.6 AI Image Generator

Draft fast on Flare. Finish on Sunburst when labels, faces, or products must stay consistent. Multi-reference image-to-image, regional mask on Sunburst, quality and complexity controls — pay per image, then animate in Studio.

Try the generator

Typical flow: explore on Flare, then regenerate the same prompt on Sunburst (or tap Upgrade below your result).

Workflow presets

Each preset sets tier, ratio, and resolution — adjust before you generate.

Reference images

Optional — up to 4 images; order matches prompt ("first image", "second image")

0/2000

~6 credits per image

Sample
Quoted poster

Built for GPT Image 2.6 workflows

Flare for speed and GIF loops. Sunburst for quoted type, multi-reference edits, and mask — without ChatGPT Plus.

Credits shown before Generate · not affiliated with OpenAI.

Vertical poster with quoted headline SUMMER DROP and product hero
01

Quoted headlines on posters

Put on-image copy in double quotes — Sunburst is tuned for tags, price lines, and short headlines.

  • Vertical 9:16 launch layouts
  • Headlines like "SUMMER DROP" stay legible
  • Sunburst · 2K — see workflow below
Split UI showing Flare draft and Sunburst finish side by side
02

Draft on Flare, finish on Sunburst

Same prompt and references — switch tier without rewriting.

  • Flare @ 1K for cheap directions
  • Sunburst @ 2K for delivery
  • Regenerate as Sunburst from your last result
Two reference thumbnails merging into one composed portrait
03

Keep / Change multi-reference

Upload up to 4 refs — image-to-image starts automatically on Sunburst.

  • Keep face, product, label in the prompt
  • Change only what you list
  • Brand-locked composites before video
Product loop with nine-frame GIF timeline UI
04

GIF loops on Flare

Format → GIF for Stories slots — default 9 frames, no video credits.

  • Flare @ 1K saves credits
  • Low complexity · simple motion
  • Credit total scales with frames
Product photo with white mask overlay on background region
05

Regional mask edit

One reference + mask — edit only the white region on Sunburst.

  • White mask = replace area
  • Product stays outside the mask
  • Pair with Keep / Change in the prompt

Workflows GPT Image 2.6 is built for

Five hands-on templates — poster first, then draft, multi-ref, mask, and GIF. Tap Use in generator to apply tier, format, and prompt.

Example 01 / 05

Output

Sunburst poster with readable quoted headline

Quoted poster (Sunburst)

Scenario
Vertical ad — headline must stay legible.
Result
High-contrast layout; copy in quotes in your prompt.
Settings
Sunburst9:162K

Prompt template

Vertical launch poster, headline "SUMMER DROP" in bold sans-serif, subline "Free shipping this week", product bottle centered below, commercial layout.

Example 02 / 05

Output

Flare draft product hero with quoted tag

Draft product hero (Flare)

Scenario
You need directions before spending Sunburst credits on a launch still.
Result
Fast 1K draft with readable tag when quoted — upgrade to Sunburst for delivery.
Settings
Flare1:11K

Prompt template

Studio packshot of a matte black earbud case on dark stone, soft rim light, shallow depth of field, paper tag that reads "BATCH 07".

Example 03 / 05

Portrait reference

Person

Product reference

Product

Output

Sunburst multi-reference brand composite

Brand-locked multi-ref (Sunburst)

Scenario
Person in one file, product in another — you need one cohesive frame before video.
Result
Single composed shot with Keep/Change style prompting.
Settings
Sunburst2 refs3:42K

Prompt template

Keep: face and skin tones from first reference; product shape and label from second reference. Change: merge into one editorial portrait, natural studio light.

Example 04 / 05

Product photo before mask edit

Reference

White = region to replace

Mask

Output

Regional mask edit on GPT Image 2.6 Sunburst

Mask regional edit (Sunburst)

Scenario
Product is right but background must change — without re-rolling the whole frame.
Result
Edit inside white mask only on Sunburst image-to-image.
Settings
Sunburst1 ref + maskAuto2K

Prompt template

Keep: product in unmasked area unchanged — label, reflections, silhouette. Change: replace white mask region with soft gradient studio backdrop.

Example 05 / 05

OutputGIF · 9f

Flare GIF social loop from GPT Image 2.6

Social GIF loop (Flare)

Scenario
Stories or ad slots need a 2–3 second loop — not full video credits.
Result
Native GIF export on Flare (9 frames by default). Simple motion reads better than busy scenes.
Settings
Flare1:11KGIFlow~52 cr

Prompt template

Minimal product hero, matte black earbud case on dark stone, soft loopable light sweep across the case, small paper tag that reads "NEW", clean background, subtle repeating motion only, no scene changes between frames.

How to generate with GPT Image 2.6

  1. Step 1

    Pick a workflow

    Flare for drafts and GIF loops; Sunburst for posters, multi-ref, and mask. Mask needs one reference + mask file.

  2. Step 2

    Set quality, complexity, and format

    Low complexity for I2I and masks; medium for posters. GIF for social loops (9 frames).

  3. Step 3

    Draft, then finish

    Generate on Flare, then Regenerate as Sunburst or switch tier before a final run.

  4. Step 4

    Download or Animate

    Open Studio with your still for image-to-video with sound.

When to choose GPT Image 2.6 on Reeldo

GPT Image 2.6 = OpenAI Flare/Sunburst, quoted type, mask, and GIF loops. Nano Banana 2.1 = fast Google multi-ref drafts. Seedream 5 = max-fidelity product and cinematic heroes.

When to choose GPT Image 2.6 on Reeldo
FeatureGPT Image 2.6Nano Banana 2.1Seedream 5
Edit consistencySunburst I2I + Keep/ChangeThinking multi-refHighest for product & cinematic heroes
Draft speedFlare @ 1KStandard @ 1K–4KSlower · worth for delivery stills
Multi-reference4 UI · up to 16 on Kie4 UI · 14 model max1 optional ref (T2I-first)
Quoted headlinesSunburst postersThinking layoutsGood for cinematic stills
Best forFlare→Sunburst · mask · GIFFast Google stills & multi-refProduct heroes & cinematic frames

Flare vs Sunburst

FeatureFlare (Draft)Sunburst (Finish)References
Default resolution1K2K—
Consistency across editsGood for draftsStronger on I2IRequired for I2I
Multi-referenceOptionalRecommendedAuto I2I
Credits / image6 / 11 / 178 / 14 / 21Same job
Best forDrafts · GIF loops @ 1KPosters · mask · finals1–4 uploads
Mask editNoYes1 ref + mask

From still to video

Finish on Sunburst, then animate — Reeldo Studio is not limited to OpenAI video models.

  1. Step 1

    Finish the still

    21:9 or 16:9 on Flare, then Sunburst if you need max detail.

  2. Step 2

    Regenerate as Sunburst

    Same prompt — one click from your last Flare result.

  3. Step 3

    Animate in Studio

    Pre-loads your frame for image-to-video.

Open image-to-video →

After you generate

Download, iterate tier, or send to Studio.

  • Download PNG/WebP/JPEG or GIF when ready.
  • Regenerate as Sunburst, or mask-edit on Sunburst with one reference.
  • Animate in Studio with native-audio video models.

FAQ

What is GPT Image 2.6 on Reeldo?

How many credits per image?

When should I use Flare vs Sunburst?

Do I need ChatGPT Plus?

GPT Image 2.6 vs Nano Banana 2.1 on Reeldo?

GPT Image 2.6 vs Qwen Image 2.1?

How many reference images?

How does mask edit work?

Transparent backgrounds?

Can I generate GIFs?

Can I make video from GPT Image stills?

Is Reeldo affiliated with OpenAI?