228 verified findings. 125 real-world hacks. $4.3B in exploit value re-discovered by AI — 4.4× our $1B goal. The AI Auditor Challenge is officially complete. Congrats to the winners! 🏆 🥇 @R4Y4N3___ 🥈 @geeknik 🥉 @abdoulrauf_jn

499 Detalks 228[CN]: The Capital-Efficient World Cup: Redefining Sports Forecasting via Yield-Bearing Prediction Markets 收益型世界杯：预测市场如何重塑体育竞猜 Time: 12, June, 2026, 20:00 (UTC+8) 🎁Rewards! We will be raffling off a limited number of exclusive Predict Fun World Cup merchandise. Join our Telegram community and participate in the AMA for a chance to win! Host: 499 Group Speaker: 阿黄 @ricardogee — APAC BD Predict Fun @predictdotfun We will explore how yield-bearing prediction markets are redefining sports forecasting with a capital-efficient approach to the World Cup.

Grok 4.5 ranks #11# out of 228 AI models with a score of 75.7/100. 🤖 • Grok 4.5 was released on July 8, 2026, and features a 500K context window with advanced reasoning capabilities. • Its strongest eligible benchmark category is Agentic, where it ranks #16#. • The model is rated among the top AI systems across the BenchLM rankings, competing with hundreds of other models. • Grok 4.5 API pricing is listed at $2 per million input tokens and $6 per million output tokens, with cached input priced at $0.30. Grok 4.5 continues to demonstrate strong performance across a broad range of AI benchmarks. Source: BenchLM

Deadlock set a new Steam concurrent peak of 228,208, according to SteamDB. The invite-only beta is at 184,122 players right now. The 24-hour peak and the all-time peak are the same number, 228,208, hit 23 hours ago. SteamDB had it third for a stretch, behind Counter-Strike 2 and Dota 2, so Valve briefly held the top three spots on its own platform. WARDOGS has since passed it. The climb lines up with the “The City Never Sleeps” update and the addition of the Rat King. The game is still not openly available.