Your product photo can move now. Same price as yesterday. OpenAI shipped ChatGPT
By Jaspinall · November 6, 2026 · Curated by George's Blog
Your product photo can move now.
Same price as yesterday.
OpenAI shipped ChatGPT Images 2.6 today. It's live in ChatGPT, Codex and the API, on every subscription tier, on web, desktop and mobile.
The headline feature is GIFs.
Give it a prompt or a photo and it makes a looping GIF. You can change the speed, edit single frames and share it.
OpenAI's own example is "turning a product shot into an eye-catching loop."
Old way: brief a motion designer, wait a few days, pay for a three-second loop.
New way: upload the packshot you already have and ask for the loop.
The second feature matters more to me.
You can now save reference images of people, pets, characters and products, then reuse them across edits without uploading again.
Anyone who has made AI product images knows the problem. The bottle changes shape. The label drifts. By edit three it's a different product.
Saved references go straight at that. Whether they hold up on your SKU is something you test, not something you assume.
The numbers OpenAI published:
1. API pricing is unchanged. $30 per 1M image output tokens.
2. Generation is 50% faster, per OpenAI.
3. The new fast API model, Flare, can return a low-quality image in about 6 seconds.
One rule before you get excited.
Amazon main and listing images still need to be static. GIFs are for everywhere else: Meta ads, email, organic social and your Shopify product page.
Here's the 3-step test I'd run this week:
1. Save your hero product shot as a reference image.
2. Ask for a short loop with one motion only. A slow turn, a pour, a lid coming off.
3. Run it in your next email or one ad set against your current static. Check click-through after a week.
If the loop wins, you got a motion asset for the price of a prompt.
If it loses, you lost ten minutes.