OpenAI revamps ChatGPT Images with faster generation, precise editing, and a new creation space

OpenAI has unveiled a major upgrade to ChatGPT’s built-in image tool, saying it now makes precise edits while keeping important details intact and can generate images up to four times faster.

The refresh, announced on 16 December 2025, is powered by the company’s new flagship image model and ships alongside a redesigned Images experience inside ChatGPT, according to the company.

A faster, more precise image model inside ChatGPT

OpenAI says the new model improves instruction-following and preserves composition, lighting, and likeness across edits, enabling practical photo retouching, believable clothing or hairstyle try‑ons, and creative style transformations that retain the essence of the original image.

The update also advances text rendering, handling denser and smaller text more reliably than the earlier release.

• Make targeted edits such as adding, subtracting, blending and transposing elements without damaging the rest of the frame.
• Create from scratch with closer adherence to multi‑step instructions and spatial relationships.
• Render more complex on‑image text, with improved fidelity over the previous version.

A new Images space for creation and reuse

Alongside the model upgrade, OpenAI is introducing a dedicated Images area in the ChatGPT sidebar for mobile and web. It offers preset filters, trending prompt ideas, and a one‑time likeness upload so users can reuse their appearance in future creations without reuploading photos each time. OpenAI positions this as a faster way to explore ideas and iterate.

How can creators get hands‑on with editing

Users can upload a picture, describe the changes they want, then select specific regions to tweak using the Select tool inside ChatGPT. The workflow supports iterative edits, from background removals to adding text or swapping objects, as per the company’s help notes.

API model, cost and enterprise rollout

For developers, the same improvements are available via the API as GPT Image 1.5, which the company says delivers more consistent logo and key‑visual preservation for brand and ecommerce use cases. OpenAI notes that image inputs and outputs are priced about twenty percent lower than the previous GPT Image 1 model, helping teams iterate more for the same budget.

On availability, OpenAI says the upgraded ChatGPT Images model is rolling out globally to all ChatGPT users and API customers, while the new Images experience inside ChatGPT is reaching most users now with Business and Enterprise access to follow.

OpenAI acknowledged that, while quality is up across many examples, results are not perfect. The company lists current limitations around certain styles, multiple‑face scenes and some multilingual use cases, indicating further iterations are planned.

Why this matters for startups and creators

The upgrade brings pro‑style controls directly into a general‑purpose chatbot that many teams already use. For Indian startups building catalogues, marketing visuals or quick A/B creative tests, stronger instruction‑following and cheaper API pricing can shorten production cycles and reduce dependency on heavy desktop workflows.

It also reflects a broader shift to AI tools that behave more like creative studios inside everyday apps.