How the contract works
This contract settles at $1 if OpenAI makes a model explicitly named GPT-6, or a clear successor to GPT-5, publicly accessible by 30 September 2026, 11:59 PM ET, including through an open beta or open waitlist. It settles at nothing if that has not happened by then. Closed betas, private access, and minor updates labeled things like GPT-5.5 do not count toward a Yes. The price at any moment reflects what buyers and sellers currently think the chance of a qualifying public release is; a contract priced at 0.30, for example, would imply a roughly three-in-ten chance, not a guarantee either way. The market resolves on 1 October 2026 based on official OpenAI announcements verified against consensus reporting from credible tech media, and a position can typically be sold before that date at whatever price the market has moved to.
Yes20%
No80%
Probability
How the price has moved
The market opened at 29% on 29 July 2026 and swung as high as 72% and as low as 28% over the following weeks, a wide range that suggests a period of active speculation, possibly around a rumor or event, that did not convert into a confirmed public release. Having tested those extremes, the price has settled at the 20% consensus level and has not moved at all in the last 24 hours or the last 7 days across 56 recorded observations. That flatness signals a market that has stopped waiting on new information and has effectively priced this as a longshot heading into the final weeks before the deadline; the move follows no single publicly reported trigger.
Analysis
Context
OpenAI has built its recent release cadence around numbered flagship models, with GPT-5 as the current public standard that ChatGPT and the API run on. A jump to GPT-6 would be the kind of announcement that dominates tech coverage worldwide, given the model's role in products from Microsoft Copilot to countless startups built on OpenAI's API.
The contract asks a narrow question: not whether OpenAI is working on a next-generation model, which it almost certainly is, but whether that model is explicitly named GPT-6 (or an unambiguous successor branding) and made accessible to the public, even through an open beta or open waitlist, by 30 September 2026. Closed, invite-only testing does not satisfy the rule. Nor does a minor version bump.
That distinction matters because AI labs frequently run private previews for months before any public rollout. The market is therefore not just pricing whether a new model exists internally, but whether OpenAI chooses, in the next several weeks, to cross the specific line into public availability.
The consensus price sits at 20% across the tracked venue, down from an opening read of 29% when the market was first recorded on 29 July 2026. In between, the price touched a high of 72% and a low of 28%, a wide range that points to at least one period of intense speculation, likely tied to a rumor, leak, or conference date, followed by a retreat once that speculation failed to produce a concrete public release. The fact that the price has since settled near the bottom of that range, unchanged over both the last 24 hours and the last 7 days, is itself informative: with 56 price observations recorded and zero net movement recently, the market has stopped reacting to new information, which usually means no new information relevant to a September deadline has arrived.
The strictness of the resolution rules is a structural reason the price sits low rather than near a coin flip. A qualifying Yes needs public accessibility, not just an internal build or a name leak, and it explicitly excludes closed betas and incremental updates. That is a high bar given how AI labs typically stage releases, with private red-teaming and safety evaluation preceding any public opening by weeks or months. If OpenAI were already running a public beta under wraps, evidence of it would likely have leaked into developer forums or media well before this stage, and the market shows no sign of pricing that in.
Volume across the tracked venue stands at $282,079, concentrated on Polymarket as the sole recorded venue. That is a modest but not negligible amount for a single-outcome tech question, and the absence of a second venue to compare against means there is no cross-market spread to read for disagreement; the 20% figure is effectively the market's one clean signal on this question right now.
Historically, major version jumps at OpenAI have not arrived on tight annual or sub-annual cycles, and a company that has just stabilized a flagship model has less incentive to fragment its user base and developer ecosystem with another full renumbering within the same year. That pattern, combined with the narrow six-week window remaining before the 30 September cutoff, is doing most of the work in keeping the price in the low-to-mid twenties rather than near even odds.
What moves the probability
No public beta signal
No open beta, waitlist, or public preview of a model called GPT-6 has been reported. Absent that, the market has little reason to price a qualifying release as likely within the remaining window.
Strict resolution wording
The rules exclude closed betas and incremental updates such as GPT-5.5, raising the bar from 'a new model exists' to 'the public can actually use it.' This pushes the probability down relative to a looser definition of release.
Short runway to the deadline
With roughly a month left before the 30 September cutoff, any qualifying release would need to be announced imminently. The shorter the window, the less time there is for a surprise launch to occur and be verified.
Release cadence after GPT-5
OpenAI's recent practice has been to consolidate around a flagship model rather than rapidly renumber. A fresh full-version jump so soon after establishing GPT-5 as the current standard would be a departure from that pattern, weighing against an early Yes.
Competitive pressure as an upside risk
Rival labs releasing major models could push OpenAI to accelerate an announcement to defend market attention. This is the main force that could move the price up quickly if it materializes, even without full public rollout.
The case for
- OpenAI has surprised the market with rapid releases before, and a sudden open beta or waitlist announcement before 30 September 2026 would satisfy the resolution rules even without a full rollout.
- Competitive pressure from rival AI labs releasing major new models could push OpenAI to announce and open access to a GPT-6-class system earlier than analysts expect.
- An open waitlist, which counts under the rules, requires far less operational readiness than a full public launch, giving OpenAI a lower bar to clear if it wants to claim the news cycle.
The case against
- No open beta, waitlist, or credible leak pointing to an imminent GPT-6 public release has surfaced as of late August 2026.
- The resolution rules exclude closed betas and minor updates, meaning OpenAI would need to cross a specific and verifiable public-access threshold, not merely announce a new model name.
- OpenAI's recent pattern has been to consolidate around GPT-5 rather than rapidly iterate to a new full version, and a six-week runway is short for a change of that scale to become public and be confirmed by credible media.
What to watch
The key date is 30 September 2026, 11:59 PM ET, the cutoff for a qualifying public release, with the market itself resolving on 1 October 2026 based on official OpenAI statements and corroborating tech media coverage. Between now and then, the things to watch are any OpenAI developer event or announcement, any leak of an open beta or waitlist for a model explicitly branded GPT-6 or an unambiguous GPT-5 successor, and any official OpenAI statement that draws a line between a minor update and a genuine next-generation release, since the rules treat those very differently.
Trade this contract
Open on Polymarket
- gas covered
