Perplexity is Redefining the Computer

Perplexity believes today’s most advanced models are already powerful enough to do extraordinary work. The problem is that chat boxes and single task agents only show a slice of that power. Even as models become more specialized, their strengths still sit in silos. Perplexity Computer is designed to pull advanced AI capabilities together into one coordinated system.

The company describes it as a general purpose digital worker. Unlike a chatbot that answers questions or an agent that completes a single task, Perplexity Computer is built to create and execute full workflows. It can run for hours or even months, breaking down a goal into tasks and subtasks, spinning up sub agents to handle research, document drafting, data processing or API calls.

Users envision an outcome, the system handles the rest.

Each task runs in an isolated compute environment with access to a real browser, filesystem and connected tools. If a subtask stalls, the system can generate additional agents to troubleshoot, search for information or write code. Work runs asynchronously, so users can step away or run multiple instances in parallel.

Under the hood, Perplexity is leaning into multi model orchestration. Instead of committing to one AI model family, it routes different tasks to whichever frontier model performs best at that job. At launch, it uses Opus 4.6 as its core reasoning engine, while delegating specialized work to models such as Gemini for research, Nano Banana for image generation,…