When will GPT Image 2 be released?

OpenAI has not officially announced a date. Based on LM Arena beta activity, the DALL-E 2/3 shutdown on May 12, 2026, and historical rollout cadence (~9 months between GPT Image 1.0 and 1.5), the best current estimate is late May to early June 2026.

Why do you think it ships before DALL-E 2/3 shut down?

DALL-E 2 and DALL-E 3 go dark on May 12, 2026. Every developer on those APIs needs a migration target. Shipping the successor after that date hands customers to Google Nano Banana 2 and ByteDance Seedream. OpenAI knows this. Historical pattern: successor models ship 2–6 weeks before deprecation cutoffs.

Could the release slip past June 2026?

Yes. Safety evaluations can add weeks unexpectedly. If OpenAI extends the DALL-E shutdown or publishes a new blog post signaling delay, this page will show the revised window within hours.

What if OpenAI names it something other than GPT Image 2?

The alert fires on any official successor to GPT Image 1.5 — whatever OpenAI names it. We're tracking the model, not the marketing name.

Will I get more than one email?

No. One email, one time, on release day. We never use your address for anything else. After the alert fires, your address is deleted from our system within 30 days unless you opt in to future alerts.

What should I use until GPT Image 2 ships?

For speed and availability today: Google Nano Banana 2 ($0.045–$0.151/image, 3–5 seconds). For lowest cost: ByteDance Seedream 5.0 Lite ($0.035/image). For ChatGPT integration: GPT Image 1.5 is still live until May 12, 2026.

Are you affiliated with OpenAI?

No. This is an independent microsite. All trademarks belong to their respective owners.

How do I know this isn't a scam to harvest emails?

Fair question. We send one email on release day — you can verify that by checking your inbox on the actual release day. If we break that promise, you have every right to mark it spam. We also never ask for more than the email itself.

Disponible hoyGoogle (Gemini)

Nano Banana 2

Google's flagship image model — our go-to for overall quality and reliable text-in-image rendering.

Fortalezas

Best overall quality and prompt adherence of the three backing models
Strong text rendering, fewer artifact fingers / tails vs prior gen
Mature image-to-image editing (closest thing to ControlNet without ComfyUI)

Limitaciones

Per-image cost is higher than Seedream Lite
Available via Gemini API; requires Google account for direct access
Artistic / stylized output is less expressive than Seedream 5.0

Fortalezas

Limitaciones

Prueba Nano Banana 2 gratis — sin registro