OpenAI Tomorrow Teaser: What We Know So Far

The short answer
On Nov 8, 2026, the official @OpenAIDevs account posted a single word, "Tomorrow," with a link. OpenAI has not said what it is launching. A Nov 3 rumor from @TokenGremlin points to up to 3 models in testing (an image model, a small efficient model and a larger, pricier one), but that is unconfirmed.
On this page (10 sections)
OpenAI's developer account posted a one-word teaser on Nov 8, 2026. @OpenAIDevs wrote "Tomorrow." with a link, at 20:04 UTC. It did not name a model, a product or a format.
That is all OpenAI has confirmed. Everything about what the post refers to is inference, and this article keeps the two apart.
What did OpenAI announce with the Tomorrow teaser?
OpenAI announced nothing beyond timing. The official @OpenAIDevs account posted "Tomorrow." on Nov 8, 2026, at 20:04 UTC, pointing to a release or event on Nov 9. The post names no model, no pricing, no API status and no capabilities.
A phone with a glowing blank screen beside a calendar page, representing a one-word launch teaser
The account is OpenAI's developer channel, which makes an API, SDK or developer tooling announcement plausible. OpenAI has not said so. The post had 294 likes and 7,989 views when captured.
For marketers, there is nothing to use yet, so keep your current production plan until OpenAI publishes specs.
What is the rumored OpenAI lineup behind the teaser?
The only specific rumor is from an independent account, not OpenAI. On Nov 3, @TokenGremlin posted that "we could be getting 3 new OpenAI models, assuming all the ones currently being tested actually make it to release."
The post lists them as:
- One image model
- One small, highly efficient model
- One larger, much more powerful and expensive model
According to the same post, all three have shown up in A/B tests or on Arena. The post text is cut off in the capture, so the rest of the claim is not available. OpenAI has not confirmed any of this, and a post dated Nov 3 does not tie these models to the Nov 8 teaser.
Is the teaser about the next OpenAI image model?
It may not be. OpenAI's homepage lists "Introducing Images 2.6," published Nov 4, 2026, with a card titled GPT-Image 2.6. That landed the day after the @TokenGremlin post and four days before the teaser.
If the rumored image model was Images 2.6, the image slot in that rumor is already filled. If OpenAI has another image model in testing, nothing in the available material says so.
We covered the launch in detail: GPT Image 2.6 Is Out: ChatGPT Images 2.6 Pricing & API. Treat that as the current OpenAI image option until a new one is announced.
What has OpenAI shipped recently?
OpenAI has shipped at a fast pace in the weeks before the teaser. Its homepage lists GPT-6.1 Terra (Oct 26, 2026) and Images 2.6 (Nov 4, 2026), alongside research and safety posts including an updated Model Spec on Oct 29.
A Zapier guide dated Oct 27, 2026 describes the cadence. It says the GPT-6 models arrived about two months after the GPT-5.6 series, which came less than three months after GPT-5.5. The same guide lists GPT Image 2.5 as OpenAI's image model at that point, before Images 2.6.
| Item | Date shown | Status |
|---|---|---|
| GPT-6.1 Terra | Oct 26, 2026 | Published by OpenAI |
| Updated Model Spec | Oct 29, 2026 | Published by OpenAI |
| Images 2.6 (GPT-Image 2.6) | Nov 4, 2026 | Published by OpenAI |
| "Tomorrow" teaser | Nov 8, 2026 | Teaser only, contents not announced |
| Rumored image, small and large models | Nov 3, 2026 (rumor) | Unconfirmed, from @TokenGremlin |
Third-party pricing tables for OpenAI's text models disagree with each other on some tiers, so check OpenAI's own pricing page for any figure before budgeting.
How to prepare your ad creative workflow for a new OpenAI model
Set up your test now, and wait for real specs before you plan around anything. A new model is only useful to an ad team if you can run it against your current winners on the same briefs, same product shots and same scoring.
A creative team desk with printed ad mockups and product shots ready for testing a new model
Practical steps:
- Pick 5 to 10 existing static or video concepts with known results as a fixed benchmark set.
- Save the exact prompts and reference images you use, so a new model gets an identical test.
- Define what counts as a win before you see outputs: text rendering accuracy, product fidelity, edit control, cost per usable asset.
- If a small efficient model arrives, test it on high volume work such as copy variants and localization. If a larger one arrives, test it on scripting and concept work.
Those last two points depend on the rumor being true. Neither model type is confirmed.
AdGraam's AI Playground lets you run many image, video and audio models side by side, which is the fastest way to compare a new release against what you already use. For static work, the Creative Generator builds ads from a product URL.
What to watch next
Watch for OpenAI's announcement on Nov 9, 2026, most likely through its blog, the @OpenAIDevs account and the developer community forum. The first things to check are the model name, whether it is available through the API on day one, and the price.
Open questions:
- Whether the release is a model, a product feature or developer tooling
- Whether it relates to the image model rumor or to Images 2.6
- Whether the small and large models from the Nov 3 rumor exist and ship together
- API pricing, rate limits and commercial use terms
- Whether it reaches free ChatGPT users or only paid tiers
This page will be updated once OpenAI publishes details.
What people are posting
Tomorrow. https://t.co/ZWapS1LIST
We could be getting 3 new OpenAI models, assuming all the ones currently being tested actually make it to release. - One image model - One small, highly efficient model - One larger, much more powerful and expensive model All three have shown up in A/B tests or on Arena at some
Key takeaways
- @OpenAIDevs posted "Tomorrow." on Nov 8, 2026 at 20:04 UTC, which points to an announcement on Nov 9.
- OpenAI has not named the product, model or format of the announcement.
- @TokenGremlin said on Nov 3 that up to 3 OpenAI models may be coming: an image model, a small efficient model and a larger, more expensive one.
- OpenAI already published "Introducing Images 2.6" on Nov 4, 2026, so any image rumor from before that date may refer to it.
- OpenAI's homepage lists Images 2.6 (Nov 4, 2026) and GPT-6.1 Terra (Oct 26, 2026) as recent launches before the teaser.
- Pricing, API availability and specs for anything new are not announced.
Compare 200+ AI video and image models side by side
Run Kling, Veo, Sora, Seedance, Flux, Ideogram and Recraft in one Playground, or turn any product URL into ready-to-launch ads.
Frequently asked
What is OpenAI launching tomorrow?
OpenAI has not said. The @OpenAIDevs account posted only "Tomorrow." with a link on Nov 8, 2026. Because it came from the developer account, an API or developer related release is plausible, but that is inference, not a confirmed fact.
Is the OpenAI teaser about a new image model?
Unknown. A Nov 3 post from @TokenGremlin listed an image model among 3 models allegedly in testing. OpenAI then published Images 2.6 on Nov 4, so that rumor may already have played out.
When will the new OpenAI release come out?
The teaser says "Tomorrow," and it was posted Nov 8, 2026 at 20:04 UTC. That suggests Nov 9, though OpenAI has not given a time or a time zone.
How much will the new OpenAI model cost?
Pricing is not announced. OpenAI has not named the model, so no per-token or per-image rates exist yet for it.
Can I use it for commercial ads?
License and usage terms for the unnamed release are not published. Check OpenAI's terms when the announcement goes live before building paid ad workflows on it.
Sources
This story is still unconfirmed: it is based on teasers and what people are posting, and the company behind it has not announced it yet. We will update this page as it develops.
More AI model news & creative guides

MiniMax H3-Image Launch: Open Weights, US Excluded
MiniMax H3-Image is out with open weights, native 4K and 16 references. See the license limits (US excluded), spec table, API pricing and ad workflow tips.

Seedream 5.5 Is Out: RGBA, 100-Round Editing
Seedream 5.5 (Dreamina Seedream 5.5) is live: native RGBA, 100-round editing, 95% text claim. Get the spec table, ad workflow tips and access details.

Claude Fable 5.5 Rumor: Leaks, Specs & Status
Claude Fable 5.5 is unannounced. See what the Azure listing, LiteLLM entry and leakers claim, plus Fable 5.1's $10/$50 pricing and what ad teams should do now.
Best AI Ad Generators for D2C Brands in 2026
Ten AI ad generators compared for D2C and e-commerce brands: what each makes, real starting prices, free options, and which suits Indian brands paying in rupees.
Break Even ROAS for COD Orders: How RTO Changes the Math
How to work out break even ROAS when a share of COD orders comes back as RTO, with worked rupee examples, a prepaid and COD mix, and why Ads Manager ROAS overstates COD.
What Indian D2C Brands Run on Meta: 2,790 Ads Analyzed
We studied 2,790 Facebook and Instagram ads from 14 Indian brands and their 747 longest-running ads: formats, angles, offers, languages and CTAs that last.
Your move
See your product as a winning ad
Turn a product URL into static ads and UGC-style videos, rebuild the competitor ads that already work, and read your Meta data to know what to make next.
AdGraam News DeskAI video and image news
The AdGraam News Desk (part of AdGraam, also searched as Adgram or Adgram AI) tracks AI image and video model launches, leaks and rumors from labs, leakers and creators, and checks each claim against its sources. Unconfirmed stories say so, and each page is updated as its story develops.