Measurement — OpenAI says Images 2.6 is 50% faster, and GPT-Image-2.6 Flare can return a low-quality image in about six seconds. Those figures are OpenAI’s.
Bill — Per-image price: not printed on the post
Door — The announcement does not name the previous image model that the 50% speed comparison uses.
On November 4, 2026, OpenAI shipped Images 2.6 in ChatGPT, Codex, and the API. OpenAI says generation is 50% faster, and GPT-Image-2.6 Flare can return an image in as little as six seconds when quality is set to low. Sunburst is the higher-quality API name. The post does not print a per-image price.
Figures from OpenAI’s November 4, 2026 Images 2.6 post.
Date
November 4, 2026
Speed
50% faster (OpenAI)
Flare latency
as little as 6 seconds at low quality
API names
GPT-Image-2.6 Flare and Sunburst
Per-image price
not printed on the post
Weekly volume
4 billion images (OpenAI)
Where it runs
ChatGPT, Codex, and the API
Quick checks
How fast is OpenAI Images 2.6?
OpenAI says Images 2.6 generates images 50% faster (post dated November 4, 2026). GPT-Image-2.6 Flare can return an image in as little as six seconds when quality is set to low. Those figures are OpenAI’s, and the post does not name the prior model used as the baseline.
What are the Images 2.6 API model names?
GPT-Image-2.6 Flare and GPT-Image-2.6 Sunburst. Flare is the faster, lower-quality setting. Sunburst is the higher-quality name for medium or high quality, with more consistency across edits on OpenAI’s account.
Does the Images 2.6 announcement print a per-image price?
No. The November 4, 2026 post points to a pricing page and does not print a per-image rate. Read that page before you budget a loop of edits. Start here: https://minerka.com/learn/.
OpenAI shipped Images 2.6 with two API names and a published low-quality latency, without a per-image price on the announcement.
What it introduces
GPT-Image-2.6 Flare and Sunburst in the API, plus GIF output and markup editing in ChatGPT and Codex.
Who benefits Anyone about to swap an image workflow onto Images 2.6, and anyone quoting “50% faster” without naming Flare, Sunburst, or the quality setting.
Try this
When a slide says the new image model is 50% faster, ask which API name, which quality setting, and faster than what. Read the API pricing page for a per-image rate before you budget. Keep the 4 billion images-per-week figure as OpenAI’s volume claim, not a quality score.
What we do not know
The announcement does not name the previous image model that the 50% speed comparison uses.
A per-image API price is not printed on the post.
No independent latency measurement is included.
Evidence
PrimaryImages 2.6 post. Speed, latency, volume, and API names are from OpenAI’s November 4, 2026 Images 2.6 post. The post points to a pricing page and does not print a per-image rate.
Independent The 50% speed claim, the six-second figure, and the weekly image volume are OpenAI’s. The post does not name the prior model used as the speed baseline.
OpenAI published “Introducing Images 2.6” on November 4, 2026. The model is available globally in ChatGPT and Codex on every subscription tier, and in the API as GPT-Image-2.6 Flare and GPT-Image-2.6 Sunburst. OpenAI says Images 2.6 generates images 50% faster than before. At quality set to low, Flare can generate an image in as little as six seconds. Sunburst is described as higher quality and more consistent across edits when quality is medium or high. The release adds GIF animation, image markup in the API, SVG export, and a complexity parameter that controls how much planning the model does. OpenAI also says people create 4 billion images per week across ChatGPT, Codex, and the API — more than half a billion a day on average. That volume figure is OpenAI’s, separate from the speed claim.
Why it matters
Teams that edit an image in a loop pay for every round. A named six-second low-quality path is useful only if you also know which API name you called and what the vendor charges per image. The announcement leaves that price off the page.
OpenAI recommends low complexity for most edits, and medium or high complexity when a scene has many constraints.
A system card is linked from the post for safety evaluations. This briefing does not restate those results.
GIF, markup, and SVG export are product surfaces. They are not a benchmark.