> ## 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-vip size Temporarily Working Again — 2K/4K Available

> The size parameter on the reverse-engineered gpt-image-2-vip is temporarily working again, producing 2K and 4K directly, though the reverse route can't guarantee permanent availability. The official-relay gpt-image-2 keeps size stable, and the docs now explain its token-based metered billing.

**2026/7/11 17:32 (UTC+8)** · Model Status · OpenAI

✅ **The `size` parameter on `gpt-image-2-vip` is temporarily working again — 2K / 4K output directly**

Following up on the [6/23 report](/en/live/2026-06/gpt-image-2-vip-size-broken-again): the `size` recognition that broke after an official rule change has recovered — passing `size` now produces 2K / 4K directly, with no workaround needed. `gpt-image-2-vip` is a reverse-engineered route, so this capability comes with **no guarantee of permanent availability** and may break again if official rules change.

Current status of each channel (pick as needed):

* Official-relay `gpt-image-2`: `size` and 2K / 4K have stayed stable throughout; metered billing by input / output tokens, with the calculation rules now documented in the [docs center](/en/api-capabilities/gpt-image-2/overview) — suited to commercial and other hard-stability scenarios
* `gpt-image-2-vip`: `size` (2K / 4K) currently works at lower cost, a good fit for personal research and learning, with availability tracking official rule changes

Thanks for following along — we keep monitoring upstream changes.

***

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