Perplexity's 19-model AI ‘Computer’
Summary
Perplexity has launched "Perplexity Computer," a new multi-model AI agent system that orchestrates 19 distinct AI models to autonomously handle complex tasks. This platform allows users to define an outcome, prompting the system to deploy sub-agents for browsing, coding, and app connectivity within a sandboxed environment. Perplexity Computer is designed for long-running operations, potentially active for months, and offers consumption-based pricing with a 10,000-credit monthly bank for its Max tier. Users can also manually select specific models for individual tasks, emphasizing flexibility as a core product feature. This release positions Perplexity as a key player in the agentic AI space, contrasting with single-model approaches from competitors like Anthropic.
Key takeaway
For CTOs and VPs of Engineering evaluating AI agent solutions, Perplexity Computer offers a compelling multi-model, sandboxed approach for long-duration tasks. Your teams can leverage its flexibility to orchestrate diverse AI capabilities, potentially reducing operational overhead and accelerating complex project completion. Consider piloting Perplexity Computer for tasks requiring extensive browsing, coding, or app integration to assess its efficiency and cost-effectiveness compared to single-model agents.
Key insights
Perplexity Computer orchestrates 19 AI models for autonomous, long-running, sandboxed agentic task execution.
Principles
- Multi-model orchestration enhances agent flexibility.
- Sandboxed environments improve autonomous agent safety.
- Consumption-based pricing aligns with usage.
Method
Users define an outcome, then the system spins up sub-agents that browse, code, and connect to apps, orchestrating 19 rival models across tasks within a sandboxed environment for extended periods.
In practice
- Automate research by defining desired outcomes.
- Use Comet to organize and rate bookmarked articles.
- Clone sales reps' voices for business scaling.
Topics
- AI Agents
- Multi-model Orchestration
- Voice AI
- AI Ethics & Governance
- Generative AI Marketing
Best for: CTO, VP of Engineering/Data, Director of AI/ML, AI Engineer, AI Product Manager, Marketing Professional
Related on AIssential
Editorial summary, takeaway, and curation by AIssential. Original article published by The Rundown AI.