OpenAI DevDay 2026: What to Expect - Confirmed, Reported, and Rumored

TL;DR
OpenAI DevDay 2026 is Tuesday, September 29 at Fort Mason in San Francisco, with Sam Altman's keynote livestreamed free at 10:00 a.m. PT. What OpenAI has confirmed, what Fortune and BleepingComputer have reported, which rumors are still unconfirmed, and what each would change for developers.
OpenAI DevDay 2026 is on Tuesday, September 29, 2026 at Fort Mason in San Francisco, and Sam Altman's opening keynote starts at 10:00 a.m. Pacific with a free livestream. That much is confirmed by OpenAI's DevDay site. What OpenAI has not confirmed is what it will announce. Fortune reports OpenAI plans to ship "a dozen or more" products at the event, including a preview of a GPT-6 Cyber model, and leak trackers have spotted an always-on assistant called "o", a faster API tier, and a $500 ChatGPT plan. This post sorts all of it into three buckets - confirmed, credibly reported, and rumored - so you know what to plan around and what to treat as noise.
Update note: This is a pre-event post, written September 28, 2026. We will update it after the keynote with what actually shipped, and mark which reports and rumors held up.
Confirmed by OpenAI#
These come from OpenAI's own pages and accounts.
| What | Detail | Source |
|---|---|---|
| Date and venue | Tuesday, September 29, 2026, Fort Mason, San Francisco | devday.openai.com |
| Opening keynote | 10:00 a.m. PT, Sam Altman | devday.openai.com |
| Livestream | The opening keynote is livestreamed free; sign up for a reminder on the DevDay site | devday.openai.com |
| Schedule | Breakouts and programming 11:15 a.m. to 3:30 p.m., closing session 4:00 p.m., all PT | devday.openai.com |
| Format | Technical sessions on APIs and tools, hands-on demos and workshops | devday.openai.com |
| Attendance | Invited attendees at $650; applications are closed | devday.openai.com |
| DevDay Exchanges | Follow-on events in Bengaluru, Tokyo, Seoul, Paris, Berlin, London, Sao Paulo, and Mexico City | devday.openai.com |
| Announcement | OpenAI calls it "our biggest event of the year" | openai.com/index/devday-2026 |
What OpenAI has not published is an agenda of announcements. Anything below this line is not confirmed.
One adjacent fact is worth knowing because it is in OpenAI's public code, not a leak: on September 25 OpenAI added an ultrafast service tier option to its official OpenAPI spec, alongside the existing fast and priority tiers. A spec change is not a launch announcement, and it says nothing about pricing or which models get it, but it is the strongest concrete signal of any item on this page.
The model lineup going into the event is also settled. GPT-6 Astra launched on September 3, and GPT-6 Sol and Luna followed on September 22 at $2/$10 and $0.10/$0.50 per million tokens. If you are expecting DevDay to be a GPT-6 launch, it already happened.
Credible Reports From Named Outlets#
These are reported by established outlets citing sources or observed evidence. OpenAI has not confirmed them.
GPT-6 Cyber and a new security product (Fortune). Fortune's senior AI reporter Emily Forlini reported on September 24, citing multiple sources familiar with the plans, that OpenAI is preparing to preview GPT-6 Cyber along with a new, yet-to-be-named product that helps customers deploy it in automated security workflows. Per the report, a limited number of Daybreak Red customers already have alpha access. Fortune later updated the story to say the launch could be within weeks rather than days, so a DevDay preview is not the same as availability.
What it would mean: if the pattern holds, this is the fourth cyber model of 2026 after GPT-5.4 Cyber, GPT-5.5 Cyber, and GPT-5.6-Cyber, and access almost certainly stays application-gated through Daybreak. The new part would be a first-party product layer on top, which moves OpenAI from selling a model to selling a patching workflow.
"A dozen or more" launches (Fortune). In the same report, a separate source told Fortune OpenAI plans to ship a dozen or more other products at DevDay, most of them not security-related, spanning enterprise and consumer. The source said OpenAI froze major launches for about two weeks, except GPT-6 Sol and Luna, to ship them together. Fortune also points to an Altman post on X teasing the event with ship emojis.
What it would mean: expect volume over a single headline. Budget time on Tuesday afternoon to read changelogs and API reference diffs, not just the keynote.
An always-on assistant called "o" (BleepingComputer). BleepingComputer reported on September 27 that OpenAI is testing an assistant called "o", which briefly appeared as a benefit on a ChatGPT Pro upgrade page, with configuration references to "o" as a display name and "-o" as an email suffix. BleepingComputer does not say it will launch at DevDay; it says it "wouldn't be surprised."
What it would mean: a persistent, provider-hosted agent with its own email identity is a consumer product first. For developers, the question is whether OpenAI exposes the same always-on runtime through the API, which would compete directly with the queue-and-cron setups many teams run today.
Rumors - Unconfirmed#
Everything in this section comes from leak trackers, contributor columns, or anonymous signals. Treat each as unconfirmed until OpenAI says otherwise.
| Rumor | Who says so | Evidence cited | Status |
|---|---|---|---|
| "o" will be announced at DevDay | TestingCatalog (September 26) | ChatGPT config strings and the Pro upgrade page; possible link to an internal "Aeon" always-on agent project | Unconfirmed |
| ChatGPT Pro Max at $500 per month with "Fastest Work and Codex" | TestingCatalog (September 24), from a post by Tibor Blaho | Unreleased subscription references; higher usage limits not confirmed | Unconfirmed |
| Ultrafast processing opens to more API users, with a Standard / Fast / Ultrafast selector in the Responses Playground | TestingCatalog (September 26) | A hidden Playground selector, plus the OpenAPI commit above | Spec change confirmed; rollout unconfirmed |
| "Managed agents" for building and running agents on OpenAI's platform | Forbes contributor Jon Markman (September 21) | Cites a TestingCatalog report on code references, not independent sourcing | Unconfirmed |
| GPT-6 Sol in the regular ChatGPT experience, desktop app changes, tighter ChatGPT, Work, and Codex integration | News9 roundup | The outlet itself labels these unverified | Unconfirmed |
What each rumor would mean if it lands:
- Ultrafast API tier. The one with the most direct developer impact. A per-request
service_tierchoice lets you pay for latency only where it matters, such as interactive agents and voice, and keep batch-style work on Standard. Watch for which GPT-6 models support it and the price multiple; Astra's existing Fast mode already runs 2x speed at 2x price. - Managed agents. If OpenAI ships a hosted agent runtime, it overlaps with what we covered in Codex, managed agents, and AWS. The decision for teams becomes build-your-own loop on the Responses API versus renting OpenAI's, with the usual trade of portability for less infrastructure.
- Pro Max. A $500 tier aimed at "fastest" Work and Codex signals that OpenAI sees long-running Codex jobs as the premium use case. If you run Codex all day, compare it against API spend, not against the $200 plans.
- "o". Mostly a ChatGPT story, unless it arrives with an API.
How to Prepare Before Tuesday#
- Pin your model ids. Launch days bring alias changes. If you call
gpt-6-solorgpt-6-astrain production, confirm you know which snapshot you are on before new versions appear. - Watch the API changelog and the OpenAPI repo, not just the stream. Last year's DevDay 2025 recap is a reminder that the developer-relevant details land in docs, not on stage.
- Have a benchmark task ready. Keep one representative agent task you can rerun against anything new, so you judge launches on your workload instead of the keynote chart.
- Check your Daybreak status if you do security work; any GPT-6 Cyber access will almost certainly flow through it.
FAQ#
When is OpenAI DevDay 2026?#
Tuesday, September 29, 2026, at Fort Mason in San Francisco. The opening keynote with Sam Altman starts at 10:00 a.m. Pacific Time.
How can I watch the DevDay 2026 keynote?#
OpenAI says the opening keynote will be livestreamed for free. You can sign up for a reminder on devday.openai.com. In-person attendance was by application and invitation, and applications are closed.
Will OpenAI announce GPT-6 at DevDay?#
GPT-6 is already out. GPT-6 Astra launched September 3, 2026, and GPT-6 Sol and Luna launched September 22. Fortune reports a GPT-6 Cyber preview may come at DevDay, but OpenAI has not confirmed it.
What is OpenAI's "o" assistant?#
It is an unannounced always-on ChatGPT assistant that BleepingComputer and TestingCatalog found references to on September 26 and 27, 2026. Whether it launches at DevDay, and whether developers get API access to it, is unconfirmed.
Is there a new faster API tier?#
OpenAI added an ultrafast service tier option to its public OpenAPI spec on September 25, 2026. Availability, pricing, and supported models have not been announced.
Sources#
Last updated: September 28, 2026
Continue Reading#
- GPT-6 Astra Release Guide - the model OpenAI brings into DevDay, with pricing and the Codex and OpenCode setup
- GPT-6 Sol vs Claude Opus 5.5 vs Grok 4.7 - where OpenAI's cheaper tiers sit in the September price war
- OpenAI Dev Day 2025: Everything Announced - last year's launches, for a sense of what DevDay usually ships
- OpenAI Codex, Managed Agents, and AWS - the hosted-agent question a DevDay launch would reopen
- OpenAI Ships GPT-5.6-Cyber Through Daybreak Red - how the last cyber model reached customers
Get the next deep dive like this in your inbox
One email a week on News and the rest of the AI dev stack. Free.
Read next on AI coding tools
GPT-6 Astra Release Guide: Benchmarks, $10/$50 Pricing, and How to Run It in Codex and OpenCode
GPT-6 Astra is OpenAI's max-capability model: 1.05M context, $10/$50 per million tokens, the first OpenAI model rated Critical for cyber, and new highs on Terminal-Bench 4.0 and OSWorld 2.0. What it is, what the benchmarks do and do not say, and the verified commands to try it.
8 min readOpenAI Dev Day 2025: Everything Announced
OpenAI is turning ChatGPT into a hub. The new Apps feature lets you access external services directly inside conversations.
4 min readGPT-6 Sol vs Claude Opus 5.5 vs Grok 4.7: The September 2026 Price War
Three frontier launches in 48 hours repriced the agentic workhorse tier: Grok 4.7 at $2/$6 (Sep 21), Claude Opus 5.5 at $4/$20 with $0.20 cache reads (Sep 22), and GPT-6 Sol at $2/$10 with Luna at $0.10/$0.50 (Sep 22). Same-day-verified rates, honest benchmark attribution, and a decision guide.
9 min readNew here? Start with
Technical content at the intersection of AI and development. Building with AI agents, Claude Code, and modern dev tools - then showing you exactly how it works.








