AI Launch Profile
OpenAI GPT Image API shutdown deadline: December 1, 2026
OpenAI announced an API lifecycle deadline for gpt-image-1-mini, gpt-image-1.5 and chatgpt-image-latest. Those model IDs are scheduled to shut down on December 1, 2026, with gpt-image-2 listed as their replacement.
At a glance
Launch Snapshot
- Company
- OpenAI
- Launch date
- June 2, 2026
- Launch type
- Not classified
- Category
- AI Developer Tools, AI Image Tools, AI Models
- Audience
- AI Artists, AI Platform Teams, Developers, Product Teams
- Pricing
- There is no separate migration fee. GPT Image 2 is billed under current API token pricing, which varies with text and image input, output, size and quality; teams should remeasure cost on representative jobs rather than carry forward old per-image assumptions.
- Free plan
- No
- API
- Yes
- Open weights/source
- No
Launch Context
Use these links to move from this record into the broader Launch Intelligence database.
Verification & Sources
- Status
- Verified
- Source links
- 4
- Freshness
- Verified July 28, 2026
- Last verified
- July 28, 2026
- Last updated
- July 28, 2026
Key source checks
Suggest a correction
Kingy AI Take
Treat December 1 as a production migration deadline, not an image-model launch. Change the pinned model ID in a test environment, then compare output fidelity, edits, masks, transparency, sizes, quality levels, moderation behavior, latency and token cost on the same asset suite. A visually plausible sample is not enough evidence for an automated creative pipeline.
Who it is for
Developers, creative-platform teams and product owners whose applications call the retiring GPT Image IDs for generation, editing, transparent assets, size or quality variants.
What feels promising
OpenAI names one replacement and maintains current generation, editing, model and pricing documentation, which makes a controlled side-by-side evaluation possible before the old endpoints disappear.
What feels unproven
Kingy did not run the retiring models against GPT Image 2. Prompt adherence, style, edit locality, alpha handling, safety responses, latency and cost may change by workload and need regression tests.
Source-backed record