Perplexity Launches ‘Perplexity Computer,’ a Multi-Model AI System That Runs Full Workflows

Perplexity on Tuesday introduced Perplexity Computer, a new system the company describes as a general-purpose “digital worker” that can create and execute full workflows using multiple frontier AI models in coordination. The product is available now to Perplexity Max subscribers, with Enterprise access planned next. Unlike a chatbot that answers questions or an agent that handles a single task, Perplexity Computer is designed to break down a user’s desired outcome into subtasks, spin up sub-agents, and run projects asynchronously for hours — or even months — across real browsers, filesystems, APIs and connected tools.

The launch marks a clear expansion beyond Perplexity’s AI search roots. After building its answer engine and later releasing Comet, its AI-native browser, the company has been layering in persistent memory, deeper research tools and task management. Perplexity Computer ties those pieces together into what it frames as a unified system capable of coordinating advanced AI models rather than relying on just one.

Here’s how it works: a user describes the outcome they want. The system decomposes that goal into tasks and subtasks, assigning sub-agents to handle research, draft documents, process data or call APIs. Each task runs in an isolated compute environment with access to a real browser and real tools. If the system encounters a problem, it can generate additional agents to troubleshoot, search for information or write code. Work runs asynchronously, allowing users to step away or run multiple instances in parallel.

Under the hood, Perplexity is leaning heavily into multi-model orchestration. At launch, the company says Opus 4.6 serves as the core reasoning engine, while specialized tasks are routed to other models: Gemini for deep research and sub-agent creation, Nano Banana for image generation, Veo 3.1 for video, Grok for lightweight speed tasks, and ChatGPT 5.2 for long-context recall and wide search. Users can also choose which model handles specific subtasks, reflecting Perplexity’s long-standing model-agnostic approach.

The strategy rests on a key premise: AI models are not commoditizing, they’re specializing. Instead of betting on a single model family, Perplexity argues the most capable system is one that can intelligently coordinate multiple frontier models and deploy each where it performs best. That orchestration layer — not the underlying model — becomes the product.

The naming is deliberate. Historically, a “computer” referred to humans who divided complex calculations among coordinated teams. Perplexity positions its system as a modern, digital version of that structure — autonomous agents dividing work, collaborating, and delivering a final output with minimal supervision.

The move also signals how AI products are evolving beyond chat interfaces. Early AI tools centered on question-and-answer interactions. Then came task-based agents. Perplexity Computer shifts the frame again, toward persistent systems that manage end-to-end workflows across time, tools and models.

For Perplexity, this is as much about positioning as functionality. The company built its brand on AI search with sourced answers. With Perplexity Computer, it’s moving into territory occupied by productivity suites, automation platforms and enterprise AI agents. By restricting access initially to Max subscribers, Perplexity is also testing the product in a premium tier before expanding to enterprise customers.

Whether users adopt the system as a true “digital worker” remains to be seen. Running long-lived, multi-model workflows introduces questions about cost control, reliability and governance. But the launch underscores a broader industry shift: the competitive edge may no longer lie in building the single best model, but in building the best system to coordinate all of them.

Perplexity Computer is live now for Max subscribers, with Enterprise Max availability coming soon — marking the company’s most ambitious step yet beyond AI search and into orchestrated, multi-model automation.

See more here:

This analysis is based on reporting from techbuzz.

Image courtesy of Perplexity.

This article was generated with AI assistance and reviewed for accuracy and quality.