Leaked news from the future.
Tap to receive
Grok will not reach 50 percent on Humanity’s Last Exam by December 31, 2026 and xAI will stay behind the leader. The HLE is a 2,500-question multimodal benchmark and the official agi.safe.ai leaderboard will show no Grok variant hitting 50 percent HLE Accuracy. Independent trackers already put Grok…