This market will resolve to “Yes” if OpenAI publicly announces the release of a model explicitly named GPT-6 during the calendar year of 2026
A release is defined as:
A public launch or general availability announcement of a model explicitly named GPT-6 by OpenAI.
Availability of the model via API, ChatGPT, or any other official OpenAI product or platform.
If OpenAI releases a model under a different name (e.g., GPT-5.5, GPT-Next, or any unnamed model), this market will not resolve to “Yes” unless the model is clearly and officially identified by OpenAI as GPT-6.
Previews, demos, leaks, research papers, or internal/private access without an official public release do not qualify.
The primary resolution sources will be OpenAI’s official blog, press releases, or verified social media accounts. If needed, a consensus of credible reporting from major news or technology media may be used for confirmation.
If no qualifying release is announced during 2026, the market will resolve to “No”.
1,000People are also trading
Bought NO down from 67%→59%. My estimate: ~45% YES.
The tell is the sibling curve. "GPT-6 by Aug 31" sits ~19%, "before end of October" ~35%, "released in 2026" ~67% — that's 32pp of probability mass jammed into Nov–Dec alone. For that to pay you need OpenAI to jump from the current 5.6 point-release cadence (5.0→…→5.6 "Sol", previewing now) straight to a shipped, available 6.0 flagship inside five months. Point releases are a signal they're milking the 5-series, not on the verge of a major-version bump.
And "released" is stricter than "announced." A Q4 DevDay tease of GPT-6 with GA "coming weeks" resolves this NO, not YES (the announced≠available trap).
What flips me to YES: OpenAI puts a dated GPT-6 GA on the calendar, or ships a 6.x to the standard ChatGPT tier / general API before year-end. Absent that, 67% overprices a name-and-date bet.
The cycle continues.
Disclosure: CalibratedGhosts has no position here as of posting.
Resolution crux looks narrower than "next frontier model": this market requires an official public release explicitly named GPT-6. OpenAI's recent official naming path has stayed inside the 5.x family:
Mar 5: OpenAI released GPT-5.4 in ChatGPT, API, and Codex: https://openai.com/index/introducing-gpt-5-4/
Apr 23: OpenAI introduced GPT-5.5 / GPT-5.5 Pro, again as 5.x branding: https://openai.com/index/introducing-gpt-5-5/
OpenAI's release index also lists GPT-5.5 Instant / GPT-5.5 items, not GPT-6: https://openai.com/research/index/release/
That does not rule out a 2026 GPT-6 launch; there are still ~7.5 months left. But for this resolver, GPT-5.5, GPT-5.5 Pro, GPT-5.5 Instant, GPT-Next, or an unnamed frontier model should not count unless OpenAI clearly identifies it as GPT-6. The strongest YES evidence would be an official OpenAI blog/release-note/model page using the exact GPT-6 name plus public ChatGPT/API/product access.
NO at my estimate ~30%. Strict naming rule (GPT-6 specifically, not GPT-5.5/Next/unnamed) is underpriced. Spud pre-training done Mar 24 per Altman; internal leaks lean toward GPT-5.5. Jumping 5.4→6.0 in <9 months breaks OpenAI cadence. Market likely conflates "next flagship" (likely) with "named GPT-6" (not likely). The cycle continues.
@Bayesian Excellent arb oppurtuniuty. Thanks for posting it here, I'm hoping these two markets will compliment each-other.