GPT Image 2.5 Sunburst
Top image model for quality and precise edits, about 21 cents a max-quality image. On a budget, use Nano Banana 2.1.
Should you use it?
Pick it if
Designers and developers who need the best edit accuracy and will pay for it.
Best for: Designers and developers who need the most accurate image edits (Image models Top 20)
Skip it if
You make lots of images on a budget (use Nano Banana 2.1, about a sixth of the price) or need results fast (use GPT Image 2.5 Flare).
You give up: About $0.21 per image at max quality; slower than Flare
The short version
- What this is
- OpenAI's highest quality image model for making and editing pictures, sold through the API since September 2026.
- Why it matters
- It is the best scoring image model right now, so it sets the quality bar; most people can get close for much less.
- Number 1 on arena.ai and Artificial Analysis for both making and editing images.
- About 21 cents per max-quality image; Nano Banana 2.1 is close behind for about 3 cents.
Good at
Our take
Sunburst is the best scoring image model on every board we checked, both for new pictures and for edits s1 s2 s3 s4. Arena.ai still marks its rank as preliminary, so it could move; we will look again when the votes settle.
The catch is cost. Artificial Analysis lists it at about $0.21 per image at max quality, roughly six times Nano Banana 2.1, which sits a few places lower s3. Flare has the same token rates and scores a little lower; OpenAI says Flare is faster s6. Pick Sunburst when the result has to be right, and Nano Banana 2.1 for volume.
Score
A score from 0 to 100 per job, against everything available today. Overall 100 across 1 job, the average of the job scores.
New to this? What the scores and labels mean
- 0 to 100
- Our score for one job, like writing or coding, compared with everything else out today. 100 is the best available today. For AI models we build it from public data, such as Arena's blind votes and Epoch AI's independent tests; apps and tools are scored on our own checklist. The same tool can score differently for different jobs, and where two likely ranges overlap, we call it a tie.
- Our verdict
- Use it: a top choice for most people. Worth trying: good, and worth a go. Depends: good for some people or one kind of job. Wait: too early to tell, or not out for everyone yet. Skip: something else does the job better.
- First look
- Just released. We say what it is and whether to bother, and give it a score once public results exist.
- No score yet
- None of the public sources we use covers it yet, or the only evidence is from a source whose numbers we can't use, so we describe it in words instead.
- Best buy
- Nothing cheaper scores as well for that job.
- Best free
- The best score for that job that costs nothing, through a free plan or by running on your own computer.
- Pay as you go
- No monthly plan: you pay for each use, usually through a developer account.
- Checked
- The day we last confirmed the score and the price against the sources. Prices change often, so check the date.
- AI model
- The engine underneath, like GPT or Claude. You use it through an app or, if you code, directly.
- Open source
- Free to download and run on your own computer, so nothing leaves your machine.
The top score of the 5 we have scored for this job.
From blind votes (Arena) · Arena data, CC BY 4.0
Early, few votesData read 20 Oct
Alternatives
Better or cheaper for images, from our scores, then the ones we'd also look at. The same kind of product comes first.
- Nano Banana 2.1FreeImage model73FreeScores 73 in images, with a free plan. Near the top of the same boards for about a sixth of the price per image. Side by side
- GPT Image 2.5 FlareImage model92Pay as you goSame token rates, slightly lower scores, and faster by OpenAI's account. Side by side
- Muse ImageImage model59FreeScores 59 for images against 100 here, at $10 per 1,000 images on Artificial Analysis, about a twentieth of Sunburst's price s3. Side by side
Evidence
Everything behind the call and the scores, folded so the page stays short. Spotted a wrong price, score or link? Report a mistake.
How we scored this
The images score is our own, built from public data: blind votes on Arena and, where they cover the job, Epoch AI's independent tests. How we combine them is our method and we don't publish it.
Public leaderboard
Reviews are drafted with help from AI models, including Anthropic's Claude. Model scores come only from outside leaderboards and independent evaluations, by a written rule.
No one paid for this review or gave us free access, and we earn nothing if you sign up. How we score
Sources 7
- s1arena.ai Text to Image leaderboard
- s2arena.ai Image Edit leaderboard (Single Image Edit)
- s3Artificial Analysis text-to-image leaderboard
- s4Artificial Analysis image editing leaderboard
- s5OpenAI API pricing
- s6OpenAI: Introducing ChatGPT Images 2.5
- s7OpenAI enterprise privacy
Something here out of date or wrong? Report a mistake and we will check it against the source.
History 2 changes
Every change to a rating, the price or our call, with the reason. Scores are against what exists on the day, so they drift down as better things ship. Older entries come from before we moved to scores.
- New: Images ratedReviewed: images (rank 1, capped while preliminary).
- Audit: added Muse Image as a much cheaper alternative. Rating re-derived, unchanged.
Use it well 4 guides on this site
Pages on this site that help with the jobs above.