“Multi-Model Is the Future”: Perplexity Doubles Down on AI Orchestration

[AIcash]

Perplexity executives argue that AI models are specializing rather than commoditizing, and their strategy reflects that belief. Instead of relying on a single large language model, Perplexity AI routes queries to whichever model performs best for a given task. Internal data shows users frequently switch between tools such as Gemini for visual outputs, Claude for software engineering, and GPT models for medical research.

The company also offers a “Model Council” feature that allows users to query multiple models simultaneously. Meanwhile, its Comet browser is set to launch on iOS next month, and a developer conference called Ask is scheduled in San Francisco. Despite online complaints about rate limits, executives maintain that no free-tier degradation has occurred — as the company increasingly tracks revenue growth over raw query volume.

[madmax]

revenue growth > query volume welcome to the adulthood phase of ai