GPT Image 2.5: Flare or Sunburst, what really changes, and what it gives on Imaginode
OpenAI released ChatGPT Images 2.5 on 8 September 2026, with two API models, GPT Image 2.5 Flare and GPT Image 2.5 Sunburst. Both have been on Imaginode since 9 September. Here is what is announced, what we measured, what each quality tier really costs in credits, and six real renders to decide between Flare and Sunburst.
Frank HoubreFounder of Imaginode
14 min read
this time, the model has actually shipped
ChatGPT Images 2.5 has been live since 8 September 2026, and the OpenAI API lists two models dated the same day: gpt-image-2.5-flare and gpt-image-2.5-sunburst.
At the end of August, we were explaining that Nano Banana 3 did not exist, and that the two anonymous models tested on LMArena in August, mona-lisa-1 and luna-lisa-alpha, smelled more of OpenAI than of Google. Well, there it is, the answer came this Tuesday 8 September. OpenAI has published ChatGPT Images 2.5, available in ChatGPT for every account, free ones included, in ChatGPT Work and in Codex, on desktop, mobile and web. Nothing officially proves that the two arena aliases were this model, but the timing matches to the day, and nobody at OpenAI has bothered to deny it.
On the developer side, the spec sheet is clear, and that is the one that matters for us. The API changelog says, under 8 September 2026: release of GPT Image 2.5 Sunburst and GPT Image 2.5 Flare for image generation and editing, through the Images API and the image generation tool of the Responses API. Both identifiers point to dated versions, gpt-image-2.5-flare-2026-09-08 and gpt-image-2.5-sunburst-2026-09-08. Two names, two models, one shared price list, we come back to that further down.
This article does what we do at every release: it separates what OpenAI announces from what we measure, it translates the token grid into credits per image, it decides between Flare and Sunburst, and it shows you real renders rather than marketing visuals. Written on release night, it was updated on 9 September: the Gateway listed both models overnight, we probed them in real requests first thing in the morning, and GPT Image 2.5 Flare and GPT Image 2.5 Sunburst have been in the Image node since this morning.
what changes, according to OpenAI and according to the first feedback
OpenAI announces latency cut by up to 50% compared with Images 2.0, more natural light, richer textures, better fidelity to reference photos and successive edits that no longer drift.
The first number is speed. OpenAI says it has cut generation latency by up to 50% compared with Images 2.0, and presents Flare as delivering better quality images than GPT Image 2 with half the wait. The Manus evaluation team, quoted in the announcement, goes further: in its tests, Flare produces its images two to four times faster than GPT Image 2, with transparent backgrounds that are finally clean thrown in, which matters to anyone making logos and interface elements.
The second axis is editing. The model is supposed to change only what you ask for, preserving the subject, the composition and the background, and above all to hold its edits across several turns of conversation. That is exactly the point we flagged as the next front in the Nano Banana article: the eighth edit that warps the face you approved at the second. If OpenAI has genuinely fixed that, it is more useful than a gain in raw beauty on a single isolated image. Transparent backgrounds and busy layouts, posters and visuals with several blocks of text, are announced as better handled too.
Now the honest caveat. OpenAI published no numbered benchmark table, no arena ranking, and the only independent feedback on release day came from a partner quoted by OpenAI. So we ran our own boards, the same ones as for every model in the catalogue: three generations in Flare, a film still, an animation frame, a poster with text, then three precise retouches of those same images in Sunburst. The six renders are below, with their exact instructions. On the main promise, the retouch that only touches what you ask for, the verdict is clear: it holds, on all three tries.
Flare or Sunburst: which one to ask for
Flare is the default choice, fast and built for volume; Sunburst is presented as the more capable of the two, cut for editing precision, with longer generation times.
OpenAI describes Flare as the everyday model: fast, good quality generation, designed for creator and social media content, product experiences, visual search, prototyping and high volume generation. Sunburst's sheet, for its part, says "our most capable model for image generation and editing", with the emphasis on edit precision, and warns of longer generation times. The official guide recommends it for the pipelines where editing precision matters most: campaign visuals ready to deliver, polished product photos.
Granted, on paper, everyone wants the more capable one. But look at the price list before choosing: both models are billed at exactly the same token rates. The difference between Flare and Sunburst is therefore not a difference in money, at least for now, it is a difference in waiting time and, if the promises hold, in fidelity on fine edits. Which gives a simple rule.
Personally, I would take Flare for everything that is still being explored: the thirty variants of a YouTube thumbnail, framing attempts, the first versions of a visual. And I would keep Sunburst for the moment when the image is approved and there are three precise edits left to make on a product photo that is going to a client. Generate fast to decide, generate slowly to deliver: it is the same logic as Nano Banana versus Nano Banana Pro, and it has never let us down.
the price: the same tokens as GPT Image 2, but two extra quality tiers
Both models cost $5 per million text input tokens, $8 per million image input tokens and $30 per million image output tokens, the exact grid of GPT Image 2; measured in real requests, they still come out three to four times cheaper than GPT Image 2 at the Standard and Max tiers, because they use fewer tokens per image.
The official price list is identical for Flare and Sunburst, and identical to that of GPT Image 2: $5 per million text input tokens ($1.25 cached), $8 per million image input tokens ($2 cached), $30 per million image output tokens. Flare's sheet says it in black and white, the token rates match those of GPT Image 2. If you read elsewhere that the price has doubled, that person is comparing with an old price list; OpenAI's pricing page, as of the evening of 8 September, shows the same numbers on both lines.
The real pricing news is elsewhere, and it is easy to miss. GPT Image 2 offered three qualities, low, medium and high. The two 2.5 models accept six on the API side: low, medium, high, xhigh, max and auto. An image is paid by the number of tokens it contains, and a higher quality tier means more tokens for the same size. On Imaginode we only show the three tiers we measured, Eco, Standard and Max, which map to low, medium and high: a setting we have not probed has no business on a button whose price is shown before you click. The two upper tiers, xhigh and max, will come once we have their real cost.
Here is what it gives, measured on 9 September, with a credit worth one cent. GPT Image 2 costs 1 credit per image in Eco, 8 in Standard, 30 in Max. GPT Image 2.5, Flare and Sunburst alike, costs 2 credits in Eco, 3 in Standard, 9 in Max. Yes, you read that right: the 2.5 Standard is cheaper than the 2 Standard, and its Max costs barely more than the 2 Standard. The reason is simple, a 2.5 image at medium quality uses roughly four times fewer tokens than a 2 image at the same setting. One detail that matters for retouching: editing an existing image costs a little more than generating at the same tier, because the input image is billed in tokens too. Our grid keeps the editing price, the highest measured, so we never sell at a loss.
A word on sizes, because the question comes up every time. The API accepts three recommended formats, 1024 x 1024, 1536 x 1024 and 1024 x 1536, and free dimensions in width x height as long as each side is a multiple of 16, the ratio stays between 1:3 and 3:1 and no edge exceeds 3840 pixels. PNG output by default, JPEG or WebP with a compression setting, transparent background on request. If you are preparing visuals for several platforms, the guide on formats and resolutions per network gives the exact sizes to ask for.
BalanceSketch, comments, templates: what ChatGPT adds, and what it is worth
In ChatGPT, Images 2.5 arrives with a sketching tool called with @Sketch, comments placed directly on an area of the image, format templates and an editing bar with version history.
On the interface side, OpenAI delivered four things alongside the model. Sketch, first: you type @Sketch in the conversation, you draw a freehand sketch, you add a style description, and the model uses it as a composition guide for the final image. Comments, next: instead of rewriting the whole prompt, you point at an area and write "make the sky more dramatic", and only that area moves. Templates, finally, for poster, flyer, merchandise and product photo, which ask structured questions before generating. And an editing toolbar: annotation, comment, background removal, eraser, resizing, with version history kept.
I will be blunt: it is rather very good, and it is not new. A sketch that guides the composition is a reference image plugged into the model's input. A comment on an area is a localised edit with a mask. A template is a structured prompt prepared in advance. They are, one by one, the building blocks of a workflow canvas, what we call nodes, and what equipped creators have been doing for a year with Flux Kontext, Nano Banana or GPT Image 2. OpenAI has just put those gestures into a chat window, which is excellent news for the hundred million people who will never open a canvas.
What the chat window still lacks, and it is structural, is repeatability. A comment on an image is done once. A chain of nodes, reference, generation, localised edit, upscale, replays across the fifty products of a catalogue by changing a single input. If you produce three visuals a month, ChatGPT with Sketch will be more than enough. If you produce three a day, the tool that counts is not the one that draws most prettily, it is the one that does the same thing again without asking you to start over. The article on reproducing the same image explains the difference in practice.
where to use it, and how on Imaginode
Four doors: ChatGPT for every account, the OpenAI API for anyone managing a key, Adobe Firefly, and Imaginode's Image node, where Flare and Sunburst have been listed since 9 September 2026 with their price shown before each generation.
The shortest door to form an opinion in ten minutes is ChatGPT. The model is rolled out to every account, free ones included, on desktop, mobile and web, and generation quotas remain those of each plan, which OpenAI still does not publish. Our comparison with ChatGPT details what those unwritten limits mean when you produce for real. Second door, the API, with the two identifiers above, a key, usage billing, and the ability to pick quality and size to the pixel. Third door, Adobe Firefly, which integrated it into its studio on day one.
And Imaginode. We plug our image and video models in through the Vercel AI Gateway catalogue, which lets us have Flux, Seedream, the three Nano Bananas, Kling or Veo behind a single credit counter. On the evening of 8 September, that catalogue did not yet list 2.5, and this article told you so rather than let you look for a missing button. Both lines appeared overnight; on the morning of the 9th, we measured the real costs per tier in real requests, set the price in credits, and published GPT Image 2.5 Flare and GPT Image 2.5 Sunburst in the Image node's list, next to GPT Image 2, which stays in place.
Concretely: add an Image node, pick Flare or Sunburst in the selector, set the quality and the format, write the prompt, and the exact price shows on the Generate button. For a retouch, plug an Image or Media node as input, describe the change in one sentence, and the model reworks your image instead of starting over. No OpenAI subscription to take, no key to manage, and the approved image stays on the canvas, ready to go to a Video or Enhance node.
what it gives: three Flare generations, three Sunburst retouches
Six renders produced on Imaginode on 9 September 2026 in Max quality: Flare generated a film still, an animation frame and a poster with text; Sunburst then changed a single element in each, and nothing else moved.
The protocol is simple, and it is the one we advise you to replay. A first Image node with Flare generates the starting image from a complete prompt, framing, light, style. A second Image node with Sunburst receives that render as input and a one-sentence instruction that starts with "change only". On the film still, the red coat turns mustard, and the woman, the rain, the signs, the reflections in the puddles and the grain stay put to the pixel. On the animation frame, the bicycle becomes a red scooter, and the fox keeps its posture, its hands on the handlebars, its raincoat; the market chalkboards, written in French, did not change by a letter.
The poster is the hardest test, because it combines accented text with a double instruction. Flare wrote "LA DERNIÈRE MARÉE" and the credit line without a single mistake on the first try, in a fifties poster typeface. Sunburst then replaced the title with "LA PREMIÈRE MARÉE" in the same font, in the same place, and turned the stormy night into a sunset, keeping the lighthouse, the cliff and the credit line. That is exactly what OpenAI promised, and what GPT Image 2 lacked: a retouch that does not recompose the image around the change.
What we keep after these six images: Flare is enough to generate, Max quality included, and it is fast, about twenty seconds per image. Sunburst takes ten seconds more and earns its place as soon as an image is approved and one correction remains to be made on it. Generate with one, correct with the other, in two nodes that follow each other: that is the chain we keep.
which model for which job, now that 2.5 is here
GPT Image 2.5 Flare at 3 credits becomes the default for generating with text in the image, Sunburst at the same price for retouching without drift, Nano Banana 2 at 10 credits stays ahead on faces in an existing photo, and Seedream 5 Lite at 5 credits on budget photorealism.
The September reflex would be to move everything to the new one. That is not far from the truth, but not for everything. GPT Image 2.5 Flare, at 3 credits in Standard, takes the place GPT Image 2 held at 8: exact text in the image, a title, a short sentence, accents, and constraint-heavy compositions. GPT Image 2 stays in the catalogue, and it still has one argument, its Eco tier at 1 credit, the cheapest in the family to rough out an idea. GPT Image 2.5 Sunburst becomes the model for precise retouching: approved image as input, a one-sentence instruction, and nothing else moves.
Nano Banana 2 at 10 credits keeps its ground: retouching a photo of a real person, removing an object, changing an outfit, holding the face intact from one image to the next. For budget photorealism, Seedream 5 Lite at 5 credits per image does things GPT Image does not do at that price, with characters that stay consistent. And for searching, trying, failing ten times, Flare in Eco at 2 credits replaces GPT Image Mini on drafts, with a base quality well above it.
The method does not change, whatever model ships next month: frame the idea on the budget model, validate the composition, and only move to the top of the range for the final version. Across ten visuals, the difference is counted in dollars. The calculator gives you the cost of a series before you launch it, and the per model pricing detail is public, tier by tier.
the model version is still not your limiting factor
What GPT Image 2.5 promises to do better, edits that no longer drift and a reference that is respected, is exactly what a node workflow guarantees by construction, whatever model is plugged into it.
Reread the list of announced improvements: the subject of the reference photo better preserved, the edit that touches only the requested area, the previous changes that hold at the next turn. Those are three flaws of the chat window, where every message puts everything back in play. In a canvas, the reference is a node that does not move, the edit is a node that receives the approved image and nothing else, and the previous version stays on the table, next to the new one. A model that drifts less is of course good news. A method where it cannot drift is better, and it already works with this summer's model.
The second lever, we repeat it because it pays more than a version change: the prompt. A three-word prompt gives a forgettable image on 2.5 just as it will on any model of 2027. A prompt that describes the framing, the light and the style, with a reference image to lock the subject, produces a deliverable image on GPT Image 2 today. The guide on writing a prompt that works takes ten minutes to read, and the one on deformed hands and illegible text fixes the two most frequent complaints, without waiting for anyone.
So yes, test GPT Image 2.5, in ChatGPT for Sketch, on Imaginode for the shown price and the node chain. But the best use of this release is to set up the reflexes that make this model useful: a locked reference, a complete prompt, a Flare node that generates and a Sunburst node that corrects, a chain you replay. Those who have them gain a generation's head start. The others will make slightly prettier images, slightly faster, with three-word prompts.