Perplexity, the AI-driven search and browser company, has launched a new agentic tool called “Perplexity Computer” designed for high-paying subscribers who need complex AI workflows handled autonomously. The tool, available exclusively to those on the $200/month Perplexity Max tier, integrates 19 different AI models and can even create subagents to tackle specific tasks.
A Shift Towards Enterprise and High-Value Users
The release of Perplexity Computer signals a strategic pivot away from mass-market growth towards a niche of “GDP-moving” users—those making critical, data-driven decisions. Company executives emphasize they are no longer focused on maximizing monthly active users (MAUs), a metric that dominated early AI hype cycles. Instead, the focus is on securing enterprise subscriptions and serving deep research needs.
This change comes after Perplexity abandoned its advertising model late last year, citing concerns that ads undermined user trust in answer accuracy. The company recognizes it can’t compete with giants like OpenAI, which boasts 800 million weekly users, and is now tailoring its offerings to a more profitable, specialized audience.
Multi-Model Orchestration as the Core Strategy
Perplexity believes the future of AI lies in combining multiple models. Rather than betting on a single “super-model,” the company argues that models are specializing —excelling at different tasks. Their software can dynamically select the best LLM (Large Language Model) for a given query, optimizing for cost and accuracy. For example, Gemini Flash is favored for visual outputs, Claude Sonnet 4.5 for software engineering, and GPT-5.1 for medical research.
The company even uses modified open-source Chinese LLMs to lower costs for certain queries, though it previously faced criticism for not disclosing this practice.
A Flawed Demo and Focus on Revenue
The launch wasn’t without hiccups: a planned press demo was canceled at the last minute due to product flaws. Despite this, Perplexity is confident in its competitive advantage—claiming its new Draco benchmark proves its deep research capabilities outperform competitors like Gemini.
Executives now prioritize revenue metrics over daily query counts, a shift that has already led to complaints from some users about stricter rate limits on free and subscription tiers. Company leadership, however, dismisses these concerns as “completely false.”
The takeaway: Perplexity is doubling down on high-end AI tools for businesses and power users, betting that specialized workflows and multi-model orchestration will be more profitable than chasing scale.
