Perplexity Computer Adds Multiagent: Up to 16 AI Agents That Cross-Check Work
Summary
- Perplexity has introduced Multiagent to its Computer platform, offering a new effort level above Ultra that coordinates up to 16 AI agents.
- The system uses a mixed-vendor default team including Claude, GPT, and Gemini models that can communicate directly and edit shared files in parallel.
- A lead agent plans the task, delegates work, and merges findings into a single response for the user.
- The feature is available exclusively on the web for Perplexity Max subscribers for complex research and multistep workflows.
Why it matters
Multiagent architecture shifts AI workflows from isolated single-model requests to collaborative, multi-model teams that can cross-check findings, enabling complex parallel tasks like financial diligence and forecasting.