Perplexity launches Computer: A multi-model AI system that can perform tasks in place of human users.
Perplexity has just introduced a new feature called Perplexity Computer , which allows multiple AI models to work together on the same task. Instead of using just a single engine, the system can flexibly utilize Gemini, Grok, or ChatGPT 5.2 depending on the nature of the job. However, the noteworthy aspect isn't the 'multi-model' nature, but the fact that the Computer doesn't just answer questions — it actually performs the work for you.
Currently, this feature has been rolled out to Perplexity Max users on the desktop web version and will soon be expanded to the Enterprise Max group.
In terms of operation, the system uses Opus 4.6 as its central inference engine. When encountering specialized tasks, it will delegate them to more suitable models. Gemini handles in-depth research tasks by creating sub-agents to delve deeper into the data. Grok is preferred for light tasks requiring high speed. Meanwhile, ChatGPT 5.2 excels in its ability to remember long contexts and perform extensive queries.
The computer also has the ability to connect with many familiar services such as Gmail, Outlook, GitHub, Slack, Notion, and Salesforce. Thanks to this, it can draft documents, build slides, send emails, manage schedules, and even run recurring tasks even when the user is offline. In other words, it's a self-operating 'to-do list' system where you simply describe your goals in natural language, and the AI will plan, execute, and report back.
Select a model for each task.
The name 'Computer' also has an interesting meaning. In the 18th century, 'computer' was a term used to refer to people who specialized in breaking down complex tasks into smaller parts. Perplexity is recreating that model in the software environment, where AI models act as a team rather than a single entity.
The major differentiating factor of Computer is its flexibility. The system isn't locked into a fixed model, and users can choose the appropriate model for each task. This is especially important given the increasing concern over token costs in the business environment. If a model consumes too many resources for a simple task, users can easily switch to a faster or more cost-effective option.
This approach transforms AI from a single-function tool into a multi-model toolkit, where each engine excels at its core strengths. Gemini can delve into research, Opus 4.6 handles complex reasoning, and ChatGPT 5.2 ensures contextual continuity in conversation. When properly coordinated, the overall system becomes significantly more powerful.
The computer also revealed a new trend in the AI industry: instead of viewing models as interchangeable products, Perplexity is betting on specialization. A single prompt can now trigger a chain of actions including research, analysis, documentation, email sending, and follow-up scheduling, all seamlessly without user intervention.
For Perplexity Max users, the Computer feature is available today. Enterprise Max will be available soon. This rollout also provides an opportunity for users to reassess the true value of coordinating multiple AI models: whether it saves time, optimizes costs, and smooths workflows.
If the system is intelligent enough to automatically select a fast model for simple tasks and a more in-depth model for complex research, the work experience will become smoother, with users having almost no need to think about the technical aspects behind it. This could be a significant step forward in how AI supports daily work.
