“For what I use AI for, perplexity is the best option by far, and I have no desire to use Chat or Claude etc. But I’m not sure that the deep research option is worth $200/year to me.”
“have had multiple occasions where my own repos aren't queries, but others. in this particular case its a public one (thank god) but i had an publicly inaccessible (Russian) one a few days ago. I'm not reporting this to perplexity bc they don't care.”
“Since Fable, I feel the models themselves cross that threshold and there is diminishing to zero marginal utility of Perplexity's own algorithms on top. I say this since I compare results from Claude, Chatgpt pro these days and the results perplexity shows is eerily similar in wording.”
“no useful models, no infrastructure, only a totally broken search engine on top of third party models that have FAR better search capacity on their own.”
“I love perplexity but I'm leaving it and changing to Claude or gpt. I don't really like in perplexity that you have to pay for a subscription and then pay for credits.”
“Perplexity had their own scanner and indexing bots, internal search engine and ranking system, video transcription, etc. agents worked on top of that, which is why they worked well.”
“Without Perplexity, if needed, 3 different AI accounts still wouldn't be as complete. They keep changing the offer, which is why I pay monthly, but the offer is always better than the alternative, so I'm still with them.”
“It's maybe not on Claude level but at least you can use it more often. And for research it depends, ChatGPT gave me some really good results and websites I never had found on my own or neither Claude or Perplexity mentioned for the same promt.”
“their browser agent is falling behind it used to be frontier research too. but after my 1y ended I see no reason to renew. I dont want the 50% off I'd rather spend my money elsewhere”
“Dude, Perplexity isn’t even a lab; their only in-house models are fine-tuned Chinese open-source models. Google DeepMind literally designs and trains its own foundational models.”
“I think Perplexity is a dark horse. Model agnostic harnesses are the future as intelligence gets commoditized. Also they were early on the “computer” thing which Grok Bot did just recently and got a ton of hype for.”
“On Perplexity, to use “Computer” (agentic work, getting real stuff done) you have to pay extra. With Claude Coworker & ChatGPT Work you can get a lot done within the $20 subscription.”
“Perplexity is different than the generalistic lumpers, in that they offer something that the single models don't; and, atop that, they offer "good use" of their range of models, to support the research you do with them, not as a full replacement.”
“Perplexity is still in its startup growth phase and are burning VC money to grow before they IPO in 2027. So they're not profitable and are subsidizing your AI costs...”
“I recommend ChatGPT Plus, as I also use Codex and find it superior to Perplexity for research—not necessarily due to the underlying model, but because of its better context management and prompt engineering. Furthermore, ChatGPT Plus offers a more versatile experience, including image generation with GPT image 2.0, a feature Perplexity lacks.”
“The catch with Perplexity Pro is they keep making the subscription worse. I had it last year on free trial and was very happy with it so I re-upped with a paid sub.”
“I use perplexity pro and will not be going back, but I exclusively use it for research and in-depth web search. Claude, ChatGPT, and even Gemini blow it out of the water for basically everything else.”
“I just canceled my Perplexity subscription for exactly the reasons you described. They seem to be pushing Computer into everything now, while Comet and the regular experience feel gutted.”
“We’ve used Google, ChatGPT, Claude, Perplexity and broader AI search to add anywhere from $20,000/month to $100,000+/month to health and wellness brands selling supplements, fitness programs, diagnostics, skincare and protocols.”