Vercel has added OpenAI's gpt-image-2 to AI Gateway, expanding the gateway beyond text and chat traffic into image generation and editing.

What changed

In its changelog, Vercel said developers can now call the model through AI Gateway by setting the model name to openai/gpt-image-2 in the AI SDK. Vercel's model page separately lists image generation support for that identifier, while OpenAI's image generation guide describes gpt-image-2 as its latest GPT Image model for creating and editing images.

Why it matters

This is not a new image model launch from OpenAI. The practical change is that teams already using Vercel's gateway can put image workloads behind the same routing and accounting layer they use for other model traffic. Vercel's docs position AI Gateway as a single endpoint for model access, budgets, monitoring, retries, and fallbacks across providers.

For teams building design tools, creative agents, or image-heavy product flows, that can remove a bit of plumbing. Instead of maintaining a separate path for image generation, they can route it through the same gateway setup they already use for text and multimodal calls.

The conservative read is that Vercel is turning AI Gateway into a broader control plane for AI workloads, not just a text-model proxy. Adding gpt-image-2 matters less as a flashy launch than as another sign that infrastructure vendors are consolidating more of the model stack behind one operational layer.