> ## Documentation Index
> Fetch the complete documentation index at: https://docs.apiyi.com/llms.txt
> Use this file to discover all available pages before exploring further.

# gpt-image-2-all latency up to ~80s · OpenAI compute-side fluctuation · speed remains its core edge

> Today gpt-image-2-all averages ~80s per image (previously ~40s with healthy pool capacity). Root cause is OpenAI compute-side fluctuation — not an IP/proxy or account issue. Size control still trails 2K gpt-image-2-vip, but raw speed remains its core advantage; still recommended.

**2026/5/23 18:16 (UTC+8)** · Model Status · OpenAI

⏱️ **`gpt-image-2-all` averaging \~80s today · \~2× the previous 40s · root cause: OpenAI compute-side fluctuation** — Heads-up: today `gpt-image-2-all` is **averaging \~80s per image** (versus \~40s previously when the account pool was healthy). We've ruled out **IP/proxy** and **account** issues — root cause is **OpenAI's compute-side fluctuation**.

💡 **Why we still recommend it**: `gpt-image-2-all` can't match the 2K tier `gpt-image-2-vip` on **size control**, but **raw speed** has always been its core edge — even at 2× the latency, it still sits in the **fast lane** versus other channels. **Fast is the edge.**

📖 Launch notes: [/en/news/gpt-image-2-all-launch](/en/news/gpt-image-2-all-launch)

***

← [Back to Live Updates](/en/live) · 📚 [Monthly Archive](/en/live/archive)