Perplexity Model Council: Querying Multiple AI Models in Parallel

What’s It About?

The AI search platform Perplexity has presented a new function called Model Council. It allows users to direct their queries to several large language models simultaneously and receive the different answers in a bundled form. The function combines the results of various AI systems and presents them in aggregated form, making a broader spectrum of perspectives and information accessible.

A so-called chair model takes on the task of bringing the various answers together and putting them into a clear form. Users should thus not only be able to grasp several viewpoints at once but also more easily recognize possible weaknesses of individual models. The function is aimed above all at professional users who depend on precise and comprehensive research results.

Background & Context

With this approach, Perplexity positions itself as a model-agnostic platform that integrates various AI systems instead of relying on a single one. The parallel use of several models is intended to combine the strengths of different systems while at the same time compensating for their respective limitations. The integrated models can include advanced systems such as GPT variants from OpenAI, Claude from Anthropic, or Gemini from Google.

The advantage lies in the time savings: users no longer have to switch between different platforms to obtain different AI perspectives. In addition, the answers are provided with concrete source references, which increases traceability. Differences between the model answers can point to different training data or focal points of the systems and help identify blind spots.

The development fits into the trend that users increasingly want to combine several AI tools for complex tasks. Perplexity tries to simplify this workflow by having the platform take over the orchestration of various models. For professional users, this can be particularly advantageous in research where different viewpoints or a higher information density are required.

What Does This Mean?

By querying several AI models in parallel, users gain a broader information base and can directly compare different perspectives
A coordinating chair model summarizes the various answers and prepares them in a structured form
Differences in the model answers make weaknesses or knowledge gaps of individual systems visible and improve the quality of research
The function saves time, as manual switching between different AI platforms is no longer necessary
Source references for the answers increase transparency and enable better verifiability of the information
The model-agnostic approach makes Perplexity more flexible in the face of the rapid development of new AI systems