Perplexity Agent API lands GLM support

Bringing GLM to the Agent API is a strategic move by Perplexity to position its managed runtime as the default orchestration layer for developers, regardless of their choice of foundation model. By integrating Z.ai's highly capable reasoning models, Perplexity bridges the gap between top-tier search capabilities and specialized agentic models.

• –**Reasoning at Scale**: Z.ai's GLM-5.2 models bring strong coding and multi-step reasoning capabilities, making them highly suited for the complex loops the Agent API is built to run.
• –**Search-Grounding for GLM**: Developers can now easily ground GLM reasoning models with Perplexity's real-time search index, solving the knowledge cutoff and hallucination issues of the models.
• –**Provider Aggregation**: Rather than managing API keys and billing for both Perplexity and Z.ai, developers get a single unified endpoint with direct provider pricing and no markups.
• –**Agentic Shift**: This update highlights the industry-wide transition from simple chat interfaces to sophisticated agent runtimes that handle tool-use, state management, and multi-model routing.