Let’s be honest: when you hear “AI research tool,” you probably think of ChatGPT. It’s the big name everyone knows. But after using Perplexity’s Deep Research feature daily for over a month now, I’m questioning why Perplexity is seen as an underdog still. Here’s why I think it’s worth your attention, even if you’re already using OpenAI’s tools.
And yes, I DID use Perplexity Deep Research to help me research and write this review about Perplexity Deep Research. Just so that’s said.
This review reflects our independent and honest opinion. We were not compensated for this review, and we covered all the costs to try the product or service ourselves.
What even is Perplexity Deep Research?
For the uninitiated, Perplexity AI is like a hybrid of Google and ChatGPT—it answers questions conversationally but cites sources like a research assistant. The Deep Research mode takes this further by acting like a PhD student who actually does their homework.
Instead of giving you a quick, surface-level answer, it spends 2–4 minutes combing through 50–100+ sources (academic papers, news articles, industry reports) to build detailed, structured reports. The wait can get a bit wait-ey at times, but it’s so, so worth it when you consider the hours and hours of manual research and drafting that I would otherwise have to do.
Why I Keep Coming Back: Standout Features
1. Breadth over bullsh*t
ChatGPT often gives confident, generic answers. Deep Research does the opposite: it shows its work. Every claim links to a source, and it prioritises recent, authoritative publications.
This doesn’t mean zero errors. I’ve caught the occasional outdated stat or misinterpreted study. But unlike ChatGPT, where fact-checking feels like a treasure hunt, this tool lets me click citations to verify sources instantly.
2. Transparency you can actually use
Most AI tools are black boxes. Deep Research shows its logic flow, like so:
3. No subscription heartburn
Using Perplexity Deep Research is free, with a five prompts per day limit. Paying Perplexity Pro subscribers, at US$20/month, get unlimited use of Deep Research.
Until yesterday, when OpenAI’s research feature would cost you US$200/month, Perplexity Deep Research, felt like a steal in comparison. Now that OpenAI has made their research feature available to ChatGPT Plus subscribers (US$20/month) the steal is slightly less steal-ey (if that’s a thing). Still, there’s no free access to the research feature in ChatGPT, so Perplexity are still ahead.
Where it stumbles (because nothing’s perfect)
1. The “Why Aren’t You Listening?!” Quirk
Deep Research sometimes ignores custom instructions from my Perplexity Spaces (workspaces where you set preferences like tone or formatting). For instance, I’ve asked it to use Australian English spelling and sentence case for headings, only to constantly get output in AmericaniZed English and headlines littered with capital letters (like 👆👆). And don’t get me started on asking for mostly paragraphs and not too many dotpoint lists…
That said, where other AI tools so far are 60% results for 30% effort, Perplexity with Deep Research is pretty close to being a perfect 80:20 tool – 80% results for 20% effort. That’s really decent. And I enjoy editing, so I enjoy working through that last 20% myself.
2. Need for speed (or lack thereof)
Responses take 2–4 minutes—lightyears in internet time. While the wait is justified by the depth, I’d love a “lite” mode for follow-up questions and quicker tasks. That said, you can wander off and return later, which takes the edge off.
3. Source control? What’s that?
While you can select whether Deep Research should look at websites, social media and/or academic sources, you can’t really do more fine-grained filtering of sources by type (e.g., “only peer-reviewed studies” or “prioritise Australian publications”). For localised topics like Australia’s AI strategy, this means sifting through irrelevant international papers.
Deep Research vs. ChatGPT: The lowdown
I’ve used both extensively. Here’s the tea:
| Scenario | Deep Research | ChatGPT Pro |
|---|---|---|
| Writing a technical blog | ✅ 80% draft-ready content with citations | ✅ 50% content, ❌ hours of fact-checking |
| Explaining niche concepts | ✅ “Here’s how blockchain oracles work, per IEEE” | ✅ “Blockchain oracles connect smart contracts” (vague) |
| Urgent deadline | ❌ 4-minute wait | ✅ Instant (but shallow) reply |
For deep dives, Deep Research is my go-to. For quick grammar checks or brainstorming? ChatGPT (you can actually also use GPT-4o within Perplexity) still wins.
Who’s This Actually For?
- Content creators who need publish-ready drafts fast.
- Students writing research papers (RIP all-nighter trauma).
- Professionals prepping reports or staying updated in their field.
It’s not for casual queries like “best pizza near me.” Save that for Google.
The verdict: Try it before defaulting to ChatGPT
Perplexity Deep Research isn’t flawless, but it does something revolutionary: it respects your time and intelligence. Instead of handing you a shiny, unverified answer, it gives you the tools to validate its work. For research-heavy tasks, that’s a game-changer—especially at $20/month with no hidden costs and no usage limits.
