跳转到主要内容
2026/11/13 13:16 (UTC+8) · Model Status · OpenAI ⚠️ All three -vip reverse models are saturated today, with frequent 429s gpt-image-2-vip, gpt-image-2.5-sunburst-vip and gpt-image-2.5-flare-vip — the three Adobe reverse (Firefly) routes — are concurrency-constrained today: requests queue and 429s are noticeably more frequent. Reverse routes are more exposed to upstream risk controls. Today a Microsoft email service outage blocked new-account registration for the account pool, leaving insufficient concurrency and amplifying the saturation. Alternative channels (pick what fits your workload):
  • Official relay gpt-image-2 / gpt-image-2.5-sunburst / gpt-image-2.5-flare: direct official connection, native 4K, billed per token
  • Official relay as fallback: keep one route primary and switch to the official relay on 429; both use the Images API, so only model changes
  • gpt-image-2.6-all: accepts a range of size values, native tier up to 1536, $0.025 per image with 2K/4K at the same price; 4K here is a super-resolution tier, not native 4K
For 429, retry with exponential backoff (5s / 15s / 45s); 429s are not billed. Content-safety errors will not change on retry — adjust the prompt instead. 📖 Related docs: -vip overview · official relay gpt-image-2 · gpt-image-2.6-all
← Back to Live Updates · 📚 Archive by Month