Grok models currently trail frontier leaders on Humanity’s Last Exam, with Grok 4.7 and 4.6 scoring 42.9–43.1% on recent independent leaderboards while Claude Opus 5.5 and Fable 5.1 reach 59–67% depending on tool use. Trader sentiment favors a Grok 45%+ outcome by year-end because xAI has delivered incremental Grok 4.x gains through 2026 and maintains aggressive release cadence, yet the 50%+ threshold remains contested given HLE’s graduate-level difficulty, rapid benchmark saturation risks, and Anthropic/OpenAI scaling advantages. Key catalysts include any Grok 5 preview, major reasoning upgrades, or December leaderboard updates that could shift implied probabilities on the 45–55% band.