Anthropic’s May 28, 2026 release of Claude Opus 4.8, which posted a leading 45.7% score on Humanity’s Last Exam under adaptive reasoning and max-effort settings, is the main factor shaping trader sentiment. The 2,500-question benchmark, developed by the Center for AI Safety and Scale AI, draws from expert contributors across mathematics, physics, biology, and other fields to test frontier large language models well beyond saturated evaluations. Claude’s rapid Opus iteration cycle has lifted results from single digits in early 2025 into the mid-40s, edging out recent Gemini 3.1 Pro Preview and GPT-5 variants. With the June 30 cutoff approaching, traders are monitoring for additional reasoning enhancements or new leaderboard entries, though the benchmark’s design keeps even top models far below human-expert levels near 90%.