Skip to main content
2026/9/21 12:20 (UTC+8) · Model Status · OpenAI ⚠️ gpt-image-2.5 official relay is short on concurrency in the Default group, while image2Enterprise runs stably The upstream concurrency quota for the default group is tight at current volumes, so requests queue or fail at peak hours. The enterprise group runs on OpenAI’s enterprise fallback route with separate capacity. Current status of both groups:
  • Default group: concurrency is tight, with queuing and a higher failure rate at peak hours
  • image2Enterprise group: ample concurrency, generating stably, at a 1.2x rate
The enterprise group’s 1.2x rate is 20% above the default group, and top-up tier bonuses offset that: the bonus rate depends on the amount of a single top-up, so at the top 20% tier what you pay matches the provider’s own pricing, while lower tiers still cost somewhat more than the default group. See Top-up Promotions for the tier table. The group is selected per Token — see Enterprise Groups for how to switch. We are adding concurrency capacity to the default group and will post an update once it is restored.
← Back to Live · 📚 Monthly archive