Research log

GPT Image 2 release status

Short answer: as of April 17, 2026, OpenAI's public image-generation docs do not list GPT Image 2 as a public model. This page helps you judge whether a signal is strong enough to keep tracking or still too weak to treat as fact.

Last updated: 2026-04-17

Timeline

2026-04-17

Status page created

The site started maintaining a dedicated GPT Image 2 research log so community observations do not leak into formal model claims.

2026-04-17

Official doc status confirmed

OpenAI's image-generation docs and model pages do not publicly list `gpt-image-2`, so it should not be presented as a released public model.

2026-04-17

Community hints archived

The user-provided duct-tape hints and the masking/gaffer/packing tape codenames from the DeepSeek share were moved into the observation layer.

Sources and confidence

SourceConfidenceStatusNote
OpenAI official image-generation docsHighVerifiedUsed to confirm the currently public model surface.
OpenAI official model pageHighVerifiedUsed to confirm the public GPT Image 1.5 listing.
Google official Nano Banana 2 / Gemini 3.1 Flash ImageHighVerifiedUsed as the official side of future comparison pages.
DeepSeek share pageMediumReadableUseful for structure, question framing, and comparison dimensions, but not treated as a final fact source.
Grok share pageLowBody unavailableBlocked by a security challenge during this review.
Gemini share pageLowBody unavailableOnly the login/navigation shell was visible in this review.

When to read this page first

  • Direct answers about release status.
  • Codenames, leak reports, and gray-release signals from the community.
  • Update dates, source tiers, and confidence changes over time.
  • Internal links into model docs and comparison pages.

What not to mistake this page for

  • Present unverified speed, parameter, or resolution claims as facts.
  • Treat community codenames as official OpenAI naming.
  • Use hype-driven or emotional headlines as research conclusions.

Research FAQ

Why make a separate status page?

Because official material, community discussion, and speculation are all mixed together right now. A separate status page makes it easier to keep formal docs clean while still tracking weak and strong signals over time.

How is this different from the model docs page?

The model docs page is the decision page. The status page is the observation log for timelines, source tiers, and confidence changes.

When should I come back to this page?

Come back when new codenames appear, shared articles circulate, official docs change, or the site starts publishing direct provider tests.