Prediction markets now price a 1 percent chance that OpenAI's next model will debut at a benchmark score of at least 1470, reflecting deep uncertainty over both the timing of GPT-6 and the meaning of the threshold itself. The contract traded $10,852 in volume over the past 24 hours and fell 80 percentage points, signaling that participants view the proposition as highly speculative. The market closes on December 31, 2026, leaving just under six months for OpenAI to ship a successor to GPT-5 and for that model to clear an evaluation bar that does not appear in any public company documentation.
OpenAI launched GPT-5 on August 7, 2025, and has stated explicitly that GPT-6 will not arrive in 2025. Industry analysts place a likely GPT-6 window in late 2026 or early 2027, but no firm release date has been announced. The company is currently expanding its GPT-5 ecosystem and reasoning-focused o-series models, including o3-pro, rather than committing publicly to a near-term numbered successor. Model release notes and help center materials describe production systems like GPT-4.1 mini, GPT-4o, o1, and o3, yet none reference a 1470 benchmark or similar scoring framework tied to future launches.
The absence of a widely cited 1470 metric in primary sources suggests the Polymarket threshold draws on internal or third-party evaluation systems rather than disclosed OpenAI performance standards. Secondary reporting and aggregator sites have promoted a first-half 2026 narrative for GPT-6, but these claims rely on speculation rather than formal company guidance. Traders face a dual uncertainty: whether OpenAI will ship a major new model before the contract expires, and whether that model will meet a scoring criterion that remains opaque to the public.
The 1 percent implied probability reflects the market's view that both conditions—a GPT-6 release in the second half of 2026 and a debut score at or above 1470—are unlikely to align. With OpenAI focused on iterating existing architectures and no confirmed benchmark framework, participants are effectively betting on undisclosed internal milestones rather than trackable public events.
