Flare is the default GPT Image 2.5 model: the same quality, editing and speed improvements as the 2.5 generation, at 50% lower latency than GPT Image 2. On ShimmerPic it costs exactly the same credits as Sunburst.
OpenAI positions Flare as "the default choice for most applications, delivering higher-quality images than GPT‑Image‑2 at 50% lower latency", suited to "creator and social content, product experiences, visual search, rapid image prototyping, and high-volume generation".
In our side-by-side runs Flare tended toward simpler compositions with less detail per element than Sunburst. If the brief is a finished campaign visual or a product hero, start with Sunburst; if you are still exploring, Flare gets you there faster at the same price.
GPT Image 2.5 was released by OpenAI on September 8, 2026 as two API models, Flare and Sunburst. Both carry the same generation improvements over GPT Image 2: sharper details, more natural lighting and richer textures, better preservation of the subjects in reference photos, more reliable multi-turn editing, and up to 50% lower latency.
Source: OpenAI, "Introducing ChatGPT Images 2.5". ShimmerPic is an independent tool and is not affiliated with OpenAI.
Same prompt on both models, one run each at 1:1, 2K, Standard. These are the Flare results; the full pairs are on the comparison page.
| Credits per image | 1K | 2K | 4K |
|---|---|---|---|
| Standard | 3 | 6 | 12 |
| High | 12 | 24 | 40 |
Credits come from one-time packs from $9 or monthly and annual plans; see pricing.
Every new account gets 6 free credits, which is two 1K images on Flare at Standard quality with a short prompt. After that, credits are bought outright or included in a monthly or annual plan. There is no unlimited tier.
Not on ShimmerPic. Both models are charged the same credits for the same resolution and quality: from 3 credits for a 1K Standard image to 40 for 4K High. Reference photos add 3 credits each and prompts over 500 bytes add a little more; the Studio quotes the exact number before you generate.
Yes. The model is a dropdown in the Studio. Keep the prompt, change the model, generate again, and both results stay in your creations.