Skip to main content
gpt-image-2-vip does not need its own skill. It shares the single gpt-image-2 series skill with gpt-image-2 (official) and gpt-image-2-all — all three use the same OpenAI Images API; only --model differs. For the full install, SKILL.md, and script, see GPT-Image-2 Series Agent Skill.

​
What this model is good for

Lockable size / 4K

Reverse Adobe line (Firefly), supports 30 size tiers (incl. 4K like 3840×2160) — controllable dimensions.

Cheapest

Flat $0.03/image, all sizes same price, no 4K surcharge.

T2I / fusion

Text-to-image and up to 16-image fusion (no quality tiers / mask).

​
How to use it in the skill

After installing the gpt-image-2 series skill, set --model to gpt-image-2.5-vip (or gpt-image-2.5-flare-vip / gpt-image-2-vip) and lock the size with --size:
To make it the default channel, add a line to gpt-image-2/.env:
Red lines: quality tiers depend on the model — the two 2.5 models take all six (xhigh / max opened 2026-09-10), gpt-image-2-vip stops at high and rejects xhigh / max; 2.5 high only equals gpt-image-2-vip medium, and 2.5 max equals its high. never send n (passing n>1 still returns 1 image but bills per count), and mask is whole-image regeneration (use the official models for precise inpainting). The skill script gates this automatically, so normal --model gpt-image-2.5-vip use is safe.Also: size is occasionally disabled upstream and forced to adaptive 1K; if so, also put the size/ratio in the prompt (e.g. “16:9”) as a fallback.