Lindy Community Icon

Join the Community

Understanding the Roles of GPT-4o Mini and Sonar Models in Perplexity | Lindy Community

Understanding the Roles of GPT-4o Mini and Sonar Models in Perplexity

·

·

Hi everyone! I have a conceptual question related to the Perplexity search step. I tried to find the answer on my own but didn’t have much luck. In this step, we can select two different models. The first dropdown (using the GPT-4o Mini model) says it "Sends a query to Perplexity AI and gets a response with real-time information and citations". The second one (using the Sonar Reasoning model) allows us to select "The Perplexity model to use for the chat. For more information see Perplexity's documentation." My question is: which of these two models is actually responsible for performing the search and the reasoning? Based on the descriptions, it seems like the first one handles both the search and reasoning, while the second one is just used to phrase or format the answer. Is that correct? I'm asking because in Perplexity’s documentation, it says that the Sonar model performs better on topic summaries and simple information retrieval, while Sonar Reasoning is better for complex analyses requiring step-by-step thinking. Since we use this Perplexity step to research leads and assign them a score, I assume Sonar Reasoning would be the better choice. But I’m not sure if in this case, it’s actually GPT-4o Mini that’s doing the search and scoring instead. Could someone please explain the difference between the two models that need to be selected in this step? What role does each one actually play? Thank you!

6 comments

Marvin A.
·
·
hey paulo! yeah that’s a bit confusing to be fair. the first model field is only used for the query. so if you set this to “manual” the model will show “(unused)” The actual perplexity search is run by the second model field. Lmk if that’s helpful!
Paulo C.
·
·
Yes, that’s helpful! Thank you! Just to confirm: we’re using the query field to “teach” this Perplexity step the rules it should follow to assign scores based on the information it finds. Based on the screenshots I uploaded, the scoring is being handled by the GPT-4o Mini model, and the actual research is being done by the Sonar Reasoning model. Is that correct? And secondly, are these the best models for this use case? With the reasoning for the score being done by GPT-4o Mini and the research being handled by Sonar Pro instead of Sonar Reasoning, for example?
Marvin A.
·
·
hmm no for this action the Perplexity model will do all of the processing. The first model field is only used when you have one of the fields on “Auto”/ “Prompt AI” I’d recommend to do the research with Perplexity and then potentially add another step for scoring (AI Agent or Thinking step is good) -> o3, Sonnet opus 4 are great models for reasoning
Paulo C.
·
·
The Perplexity step uses a text input that’s set to AI prompt mode. In this prompt, we give instructions based on the information received in the previous step, define what it should look for, and then provide the “rules” it should follow to assign a score based on what it finds. Given that it's in AI prompt mode and using GPT-4o Mini, does that change your answer in any way? Either way, I’ll test the approach you recommended by letting Perplexity handle the research and assigning the score in the next step. Thanks so much!
Paulo C.
·
·
Forgot to ask something but if I build it the way you recommended, the model I should use for the perplexity step that is only going to do the research is Sonar or Sonar pro, correct? Because the score is being given by the ai agent, so the perplexity step isn't doing any reasoning
Marvin A.
·
·
yeah Perplexity does it’s own reasoning/thinking. -> you take that result and give it to an AI agent for scoring 🙂