Skip to main content
gpt-image-2.5-flare-vip and gpt-image-2.5-sunburst-vip are temporarily unavailable due to limited upstream capacity. If you need GPT Image 2.5, switch to the official-forward models gpt-image-2.5-flare / gpt-image-2.5-sunburst, or use gpt-image-2.5-web.
  • Published: September 16, 2026
  • Last checked: September 16, 2026 (LaoZhang API operations notice)
  • Current status: Active; both GPT Image 2.5 -vip models are unavailable, while official-forward and -web routes are working

​
Affected models and available alternatives

This incident affects only the two GPT Image 2.5 -vip models listed above. Users of official-forward models, gpt-image-2.5-web, and gpt-image-2-vip do not need to switch routes because of this notice.

​
What to do

  1. Pause new requests to both 2.5 -vip models and move jobs that need to continue to an available route.
  2. For explicit Flare / Sunburst selection or 2K / 4K output: choose an unsuffixed official-forward model. Check official-forward permissions and usage-based billing in token management, then follow the GPT Image 2.5 integration guide.
  3. For GPT Image 2.5 with per-generation billing: use gpt-image-2.5-web. It accepts size, but cannot select Flare / Sunburst. Do not assume that every parameter from the two -vip models carries over; verify the parameters and output you need with your current account.
  4. If GPT Image 2 is sufficient: use gpt-image-2-vip and follow the GPT Image 2 guide.
Before resuming batch jobs, run a small-volume request, open the resulting image, check its dimensions, and review the model, usage, and charge in call logs. Official forwarding uses usage-based billing; check the console model and pricing page before switching.

​
Frequently asked questions

​
Is all GPT Image 2.5 access unavailable?

No. Only gpt-image-2.5-flare-vip and gpt-image-2.5-sunburst-vip are affected. The unsuffixed official-forward models and gpt-image-2.5-web are working.

​
Can the web route select Flare or Sunburst?

No. gpt-image-2.5-web supports size, but cannot manually select Flare / Sunburst. Use the corresponding official-forward model if you need a specific variant.

​
Will the replacement still use per-generation billing?

gpt-image-2.5-web and gpt-image-2-vip use per-generation billing. Official-forward gpt-image-2.5-flare / gpt-image-2.5-sunburst use usage-based billing. Check both the token configuration and the target model’s price.

​
When will the two 2.5 -vip models return?

There is no confirmed recovery time. We will update this notice and the latest announcements when upstream capacity is restored. Wait for a recovery notice before switching back.