Perplexity, a popular AI startup, hasn’t been in the news much over the past few months. Now, it’s back with a new product called Perplexity Computer. The idea is to unify different AI capabilities into a single system that users can rely on to get tasks done.
The Perplexity team wrote the following about the launch:
The best AI models in the world are already capable of extraordinary work. What they’ve lacked is a system that can orchestrate capability across tasks, tools, and time.
Perplexity Computer is essentially a general-purpose agent that can create and execute end-to-end workflows based on a user’s query. When a user enters a prompt, Perplexity Computer breaks it into tasks and subtasks, then spawns multiple sub-agents to carry out the work. For example, these sub-agents can do web research, generate documents, process data, and more. Users can also run multiple Perplexity Computers in parallel to handle different tasks at the same time.
Perplexity says each task runs in an isolated compute environment with access to a real filesystem, a real browser, and real tool integrations. The system is built on models from multiple companies, reflecting the reality that different models excel at different things. To take advantage of the best model for each job, Perplexity is taking a model-agnostic approach.
For now, Perplexity Computer uses Anthropic’s Opus 4.6 as its core reasoning engine to orchestrate sub-agents. It uses Gemini for deep research, Nano Banana for images, Veo 3.1 for video, Grok for speed on lightweight tasks, and ChatGPT 5.2 for long-context recall and wide search. This model-agnostic setup should allow Perplexity to swap in newer models as they improve.
Perplexity Computer is now available to Perplexity Max subscribers, and it will roll out to Enterprise Max users in the coming weeks.