Two days ago, this publication reported that prediction markets were giving OpenAI's GPT-5.5, codenamed "Spud," a 75% chance of launching on April 23. That number has climbed. As of Wednesday morning, Polymarket assigns an 86% implied probability to a release today.

More telling than the number itself is the pattern behind it. Anonymous whale bets from newly created wallets have piled in on April 23 in the past 48 hours. Polymarket analysts note that these wallet patterns mirror the ones that correctly called the Claude Opus 4.7 release date weeks before Anthropic announced it. The remaining 10.1% odds sit on April 24, a hedge against the typical last-minute delays that plague AI model rollouts.

What We Know About Spud

OpenAI completed pretraining for GPT-5.5 on March 24, 2026, per CEO Sam Altman's public statements. The model has been in safety review since. Early leaked benchmarks, first surfaced by Geeky Gadgets and World of AI, show Spud outperforming Claude Opus 4.7 and Google's Gemini 3.1 in front-end design automation, SVG generation, and agentic workflow tasks.

Leaked demonstrations include 3D world generation using Three.js (a functional recreation of Monica's apartment from Friends), a playable Minecraft replica with real-time terrain modification, and a fully interactive solar system model. These capabilities suggest a significant step up in spatial reasoning and code generation compared to GPT-5.4, which launched on March 5.

There is confusion in the market about naming. Some outlets refer to the model as GPT-6, others as GPT-5.5, and OpenAI has not publicly clarified. FindSkill.ai's analysis describes it as a "bridge model" between GPT-5.4 and a future GPT-6, designed for faster response times and improved token efficiency rather than a generational leap in raw capability.

Why Today Matters for OpenAI

OpenAI needs this launch. Anthropic just posted $30 billion in annualized revenue, surpassing OpenAI's $25 billion for the first time. Claude Opus 4.7 dominates coding benchmarks. Ramp data shows Anthropic now claims 37% of trackable enterprise AI spending versus OpenAI's 33%. Three top executives left OpenAI on a single day last week.

A strong Spud launch would recapture the narrative at a critical moment. OpenAI's $122 billion funding round in March valued the company at $852 billion. That valuation needs to hold through an IPO process. Investors are watching whether OpenAI can compete with Anthropic on the enterprise metrics that actually drive revenue, not just consumer chatbot traffic.

The Prediction Market Track Record

Polymarket's AI release markets have been remarkably accurate in 2026. The platform correctly predicted Claude Opus 4.7's release timing weeks in advance. Whale bettors with apparent insider knowledge have used the same playbook repeatedly: fresh wallets, large directional bets, concentrated timing.

If the pattern holds, today is the day. If not, the 10.1% April 24 odds suggest the delay would be measured in hours, not weeks. Either way, the prediction market signal is the strongest it has been since the contract opened.

We will update this story when OpenAI makes its announcement. If Spud drops today, expect a full technical breakdown and competitive analysis in the next cycle.

Sources: Polymarket prediction data, Geeky Gadgets (leaked benchmarks), FindSkill.ai (model positioning analysis), World of AI (demo leaks).