2 hours ago
1
Anthropic, Google, and GitHub’s AI coding agents were compromised through prompt injection attacks. The “top AI model by June 2026” market for Google holds at 16.5% YES, down from 18% yesterday, while the “best AI model by end of April” market for Anthropic shows reduced confidence.
The Anthropic market hasn’t seen a significant volume change, but the security breach is likely to weigh on sentiment. With only 14 days until resolution, any delay in Anthropic’s model improvements could hurt its ranking.
The June 2026 Google AI market dropped 2 points after the news, even though the vulnerabilities were promptly patched. The market is thin: just $6 is enough to move the price 5 points, which says more about low participation than about broad trader conviction.
Actual USDC trading volume is low, with $59 daily for the Google market. The 2-point price drop occurred at 11:07 AM, likely from a single modest trade. At this liquidity level, one order can swing prices, so the move reflects individual positioning rather than consensus.
The hack puts a concrete number on security risk for AI model rankings. A YES share at 16¢ pays $1 if Google tops the leaderboard by June, a 6.25x return. Traders pricing these contracts need to weigh how security vulnerabilities affect both model performance and public perception of these companies.
Watch for announcements from Anthropic and Google about model updates or security fixes. Any news on either front could shift odds sharply before the end-of-April and June resolution dates.
Get prediction market intelligence as a structured API feed. Early access waitlist.
English (US) · 