Perplexity's internal WANDR benchmark, measuring agentic research capabilities, shows Grok 4.5 outperforming all other configurations, including Perplexity's own GLM 5.2 post-trained model. Grok 4.5 is now available as an orchestrator model for Perplexity's Consumer Pro and Max subscribers, offering its superior performance at approximately half the cost of Claude Opus 4.8.