Perplexity has unveiled Perplexity Computer, a multi-model orchestration system that distributes tasks across 19 different AI models, positioning the company among the first to make model flexibility a core feature of its platform.
The details
Users describe the desired outcome, and the system launches specialized sub-agents that can browse the web, write code, connect to apps, and complete tasks autonomously.
Each job operates inside its own sandbox while coordinating multiple models across the workflow. The platform can combine competing models in a single task chain and is designed to run continuously for extended periods, potentially for months at a time.
CEO Aravind Srinivas also took a direct jab at Anthropic, writing that “the biggest weakness of Claude is that it only coworks with Claude.”
Pricing follows a consumption model. The Max tier includes a monthly bank of 10,000 credits, and users can also choose which model they want handling specific tasks.
Why important?
Multi-model choice has slowly been entering AI products, especially in creative tools. Perplexity Computer is one of the first major attempts to integrate that flexibility into an agent system capable of running long term autonomous workflows, while keeping each process contained within a sandboxed environment for safety and control.
Sources: