X @Elon Musk

Model Performance and Benchmarking - Grok 4.5 ranks 4th on the GDPval-AA v2 leaderboard with an Elo rating of 1543 [1] - Grok 4.5 performance in real-world agentic knowledge work tasks currently trails behind the latest Claude models from Anthropic [1] Cost Efficiency and Market Positioning - Grok 4.5 demonstrates high cost-efficiency with a processing cost of $0.49 per GDPval task [1] - The model maintains a competitive position on the Pareto frontier regarding the balance of performance and cost [1]