I Ran 10 Prompts to See Who Wins

Everybody’s evaluating AI chatbots — however what occurs when certainly one of them shouldn’t be a chatbot in any respect? 

That’s what instantly intrigued me about Perplexity AI. It manufacturers itself as an ‘AI-powered reply engine’ — a citation-rich, clever various to Google. But, in apply, it typically seems like a chatbot, delivering solutions immediately, albeit with a powerful analysis spine.

I’ve been utilizing it because it first launched in late 2022, proper across the time ChatGPT exploded onto the scene. For sure, I discovered myself always switching between the 2, testing concepts, writing drafts, and digging into analysis. And whereas they appear to serve completely different functions on paper, in actuality, there’s lots of overlap.

So I lastly did it: Perplexity vs ChatGPT, face to face. Identical prompts. Identical duties. Identical expectations. From fact-checking to content material creation, I wished to see which one truly delivers extra worth whenever you’re deep within the circulation of labor.

And right here’s what occurred, all with G2 knowledge to again it up

Notice: Each OpenAI and Perplexity AI incessantly roll out new updates to those AI chatbots. The main points under mirror essentially the most present capabilities as of April 2025 however could change over time.

• Positioning and function: From what I’ve seen, ChatGPT is clearly designed as a general-purpose AI chatbot — inventive, conversational, and customizable. Perplexity, alternatively, positions itself as an “reply engine” constructed for quick, correct, and sourced responses. It feels extra like a wise, AI-powered various to Google than a basic chatbot.
• Interface expertise: ChatGPT seems like a chat app at its core. It’s extremely succesful and designed for longer, multi-turn conversations. With chat historical past, customized GPTs (tailor-made utilizing directions, information, or features), and built-in instruments like shopping and picture era, it creates a versatile house for creativity and problem-solving. Perplexity, alternatively, leans into its identification as a search-first device. Its interface resembles a analysis engine with a chat overlay optimized for quick, citation-rich responses. It additionally contains options like Uncover for trending subjects and Areas to avoid wasting solutions, manage analysis, or construct light-weight customized AI assistants.
• AI fashions and processing energy: ChatGPT runs on OpenAI’s GPT fashions, whereas Perplexity stands out for providing quite a lot of LLMs from a number of suppliers, together with Claude 3.5 Sonnet, GPT-4o, Gemini Flash, and extra. It additionally options its personal fashions like Sonar and R1 1776, giving customers larger flexibility and management.
• Quotation-first vs. chat-first: Perplexity all the time exhibits sources. Citations are constructed into each reply. With ChatGPT, you’ll solely get sources when utilizing the net shopping device (and even then, they’re much less distinguished.)
  
• Reminiscence and continuity: ChatGPT helps reminiscence throughout periods, which implies it could keep in mind preferences, previous prompts, or context I’ve shared (tremendous helpful for ongoing workflows.) Perplexity doesn’t have long-term reminiscence, no less than from what I’ve noticed. So, every session typically begins contemporary.

Perplexity and ChatGPT: The similarities

Regardless of the branding and options, they nonetheless have loads in widespread on the subject of getting stuff carried out.

• Textual content era: Each instruments are nice at producing clear, human-like responses to questions, prompts, and inventive duties. Whether or not I’m summarizing an article, drafting a weblog intro, or rephrasing one thing for tone, each ChatGPT and Perplexity ship coherent, context-aware output.
• Coding: Whereas ChatGPT has the sting for extra superior dev duties, Perplexity nonetheless holds its personal for fast code explanations, syntax assist, and debugging recommendations. I’ve used each for all the pieces from fundamental HTML and Python snippets to exploring new frameworks.
• Voice interactions: Each platforms now assist voice chats. I’ve used ChatGPT’s voice mode for informal conversations or fast prompts on the go, and Perplexity just lately rolled out voice capabilities, too. It’s nice after I need quick solutions with out typing.
• Multimodal capabilities: Every platform helps multimodal enter in numerous methods. ChatGPT lets me add pictures and have the mannequin describe, analyze, or interpret them. Perplexity isn’t natively constructed for picture era, however it could extract insights from internet visible content material and generate pictures with the paid plan.
• File evaluation: Each instruments allow you to add information, like PDFs, docs, or decks, for summarization and Q&A. I’ve used them to extract key insights from dense analysis papers in seconds. Perplexity helps plain textual content, code, PDFs, and pictures as much as 25MB, with 3 uploads per day at no cost customers and as much as 4 information without delay. ChatGPT helps codecs like DOCX, PDF, TXT, PPTX, CSV, and XLSX (as much as 512MB every), with free customers once more restricted to a few uploads every day.
• Net search and deep analysis: Each ChatGPT (through SearchGPT) and Perplexity present real-time internet entry, however their kinds differ—ChatGPT summarizes, whereas Perplexity highlights sources and citations by default. Each additionally supply a Deep Analysis characteristic that pulls from a number of internet sources to generate structured, in-depth reviews. I’ve discovered it nice for tackling advanced subjects or multi-layered questions.
• Short-term chats and sharing: Each instruments assist short-term periods and make sharing straightforward. I typically share Perplexity threads by downloading them as PDFs or textual content information, whereas ChatGPT gives shareable chat hyperlinks.
• Customized AI assistants:  Whereas ChatGPT has “customized GPTs” and Perplexity has “Areas,” each let me construct purpose-specific AI instruments. For instance, I might create a writing assistant that remembers my most popular type in a customized GPT, or a Perplexity House centered on monitoring inventory market information with particular sources. Each platforms make it straightforward to customise how the AI responds and remembers context (inside that session).

Evaluating specs is one factor. However how do Perplexity and ChatGPT maintain up in apply? Right here’s how I put them to the take a look at.

Disclaimer: AI responses could differ based mostly on phrasing, session historical past, and system updates for a similar prompts. These outcomes mirror the fashions’ capabilities on the time of testing. Additionally, this overview is a person opinion and doesn’t mirror G2’s place concerning the talked about software program’s likes and dislikes.

The primary problem concerned summarizing. I instructed each ChatGPT and Perplexity to extract the important thing data from a G2 article detailing the rising adoption of Canva by non-designers, presenting it in precisely three bullet factors and beneath 50 phrases.

Perplexity’s response to the summarization immediate

Instantly, I seen a distinction in how they approached the duty. Perplexity saved issues clear and direct. Its bullets have been quick, skimmable, and stayed true to the temporary.

ChatGPT’s response to the summarization immediate

ChatGPT, in the meantime, added extra element and depth, even bringing in G2 overview knowledge, which was spectacular however went a bit overboard for a decent abstract. It felt extra like a mini editorial than a TL; DR.

Whereas each have been correct, Perplexity’s model was simpler to make use of at a look. So if I wanted one thing polished for a write-up, I would lean on ChatGPT. However for fast, high-impact summaries, particularly beneath phrase limits, Perplexity did a greater job.

Winner: Perplexity

2. Content material creation

Transferring on to AI content material creation, a identified power, I wished to see how Perplexity and ChatGPT would carry out beneath the strain of a full advertising push.

So, I gave them each a fairly complete single immediate, asking for product descriptions, catchy taglines, social media posts for various platforms, e-mail topic traces to attract individuals in, and even a brief script for a video advert. Mainly, the entire 9 yards of a advertising marketing campaign!

Each ChatGPT and Perplexity dealt with it very well. The outputs have been polished, diverse, and genuinely usable. Apparently, the concepts they got here up with have been fairly comparable throughout each instruments, which made the comparability really feel even fairer.

Perplexity’s output was sturdy. Its tone shifted properly between platforms — playful on Instagram, easy on e-mail, and visible on video. 

Perplexity’s response to content material creation immediate

I discovered its tagline punchier than ChatGPT’s. Its copy didn’t really feel templated, and I appreciated that it didn’t want a lot tweaking. I particularly appreciated how naturally it dealt with completely different codecs with out dropping model voice.

Perplexity’s response to content material creation immediate

ChatGPT made the content material really feel able to drop right into a model doc or marketing campaign deck. It additionally provided extra hashtag choices for social media posts, which is useful in the event you’re attempting to cowl a number of angles or faucet into completely different traits. The tone throughout codecs was constant, and I didn’t spot any weaknesses in its method.

ChatGPT’s response to content material creation immediate

Backside line: each instruments carried out impressively right here. I didn’t really feel like both one lagged. If I had to decide on, I’d say it’s a tie; ChatGPT wins on construction and extras (like hashtag protection), whereas Perplexity stands out for its fluid tone and plug-and-play readiness.

Winner: Cut up verdict.

3. Inventive writing

I actually wished to see how effectively these instruments might break free from formulaic outputs and truly inform a narrative with temper, pacing, and a twist. I gave each ChatGPT and Perplexity a sci-fi immediate with a number of must-have components: a mysterious sign, a sentient AI, and a reality-bending reveal — all inside 300 phrases.

Proper off the bat, ChatGPT stood out for together with a title, “Whispers of the Wanderer,” which immediately set the tone. Its story had ambiance, rigidity, and a cinematic really feel. The pacing was tight, the language vivid (particularly these descriptions of the nebula and the glitching hologram), and the ending twist, “You’re the sign, ”landed completely.

ChatGPT’s story “Whispers of Wanderer” for the inventive writing job

Perplexity’s take was additionally sturdy. It constructed a special form of temper, which was extra philosophical and nearly dreamlike. The narrative had a softer tone however nonetheless hit the important thing components. The ultimate line, “Actuality shouldn’t be what you see, however what you’re allowed to see,” was a strong nearer. It leaned barely extra summary, however I appreciated that it took a special stylistic route than ChatGPT’s model.

Perplexity’s response to my inventive writing immediate

Each tales had depth and stable character voice and used the weather I requested for. So total? One other sturdy displaying from each. 

Winner: Cut up

4. Coding  

Full disclosure: I’m not a developer. However I do know that coding is a significant benchmark for AI efficiency, particularly on the subject of real-world use. For this take a look at, I requested each ChatGPT and Perplexity to construct a easy password generator utilizing HTML and JavaScript. I wished to get a working answer with clear code and a user-friendly interface.

And this spherical? ChatGPT swept it. The code it generated labored completely on the primary strive. The interface was clear and intuitive, and the device did precisely what it promised — no hiccups. At the same time as a non-dev, I might perceive what the code was doing, and the general setup appeared polished sufficient to drop right into a newbie undertaking or a fast demo. I appreciated that it had styled the UI additionally higher with the lock emoji and colourful buttons.

ChatGPT’s password generator

Perplexity, alternatively, produced a principally useful model, however the clipboard copy didn’t work. That may look like a small element, but it surely made the entire expertise really feel much less full. The UI additionally wasn’t fairly as refined. It did the job, however lacked the little touches that made ChatGPT’s model really feel extra usable and polished.

Perplexity’s password generator

Winner: ChatGPT

5. Picture era

Subsequent, I wished to check one thing a bit of extra visible — picture era. We’ve all seen AI-generated artwork floating round on-line, however I used to be curious to see how effectively these instruments might deal with one thing grounded and real looking: a inventory photograph of a small enterprise proprietor. It’s the form of picture entrepreneurs, content material creators, and small groups always want. However producing one that truly seems to be plausible? That’s an actual problem.

It’s price noting that Perplexity lets solely Professional customers generate pictures as a part of their workflow, and I attempted it with each Flux and DALL·E 3 on the platform. 

ChatGPT, utilizing GPT-4o, gave me what felt like the most effective total interpretation. The setting appeared like a comfy boutique, full with a mixture of merchandise — garments, equipment, and a heat, trendy vibe. It checked a lot of the containers in a balanced, visually clear means.  

And given all of the current buzz round GPT-4o’s picture era capabilities, having the ability to create all the pieces from Ghibli-style artwork to ultra-realistic images, it really delivered right here. The picture appeared polished, pure, and really usable for real-world content material.

Picture generated with ChatGPT

Perplexity’s picture era expertise was a bit much less intuitive. I needed to dig by means of the interface to really discover the choice because it wasn’t entrance and heart like in ChatGPT. However as soon as I received there, the outcomes have been stable. The model generated by means of Flux leaned into the ambiance greater than element. It didn’t cowl each side of the immediate, but it surely received the comfortable, inviting temper proper. If ambiance was the purpose, this was a win.

The DALL·E 3 model from Perplexity was essentially the most product-specific. It zoomed in on the handmade and textile angle, giving off severe artisanal store vibes. Consider carefully stacked materials, stitching instruments, and a workspace that feels curated and genuine. If I have been selling a craft-focused small enterprise, this might be the one I’d attain for.

Picture generated with Perplexity

So, which device was higher? Truthfully, it is determined by what you’re prioritizing. ChatGPT was the most effective all-rounder, Flux delivered on vibe, and DALL·E 3 on product specificity. Completely different flavors, all spectacular — however ChatGPT positively felt essentially the most seamless to make use of.

Winner: ChatGPT

6. Picture evaluation 

For picture evaluation, I actually wished to push Perplexity and ChatGPT a bit additional. As a substitute of a easy image, I gave them two distinct sorts of visuals: an infographic about AI adoption and a handwritten poem. And actually? Each instruments did surprisingly effectively.

Perplexity provided clear summaries for each pictures, highlighting key traits within the infographic and decoding the poem’s visible components with ease. It pulled out crucial percentages, provided insights about design decisions, and interpreted the poem with emotional nuance. No main crimson flags there!

Perplexity’s response to my picture evaluation immediate

ChatGPT additionally did a stable. However it was far more structured than Perplexity’s with subheaders. For the handwritten poem, it went the additional mile and absolutely transcribed the poem, which I discovered tremendous useful. That little little bit of added construction made it simpler to skim.

ChatGPT’s response to my picture evaluation immediate

First, the infographic. ChatGPT gave me a brilliant well-structured abstract, hitting all the important thing statistics, traits, and conclusions. It even gave me some ideas on the visible design, which was a pleasant contact.

If I needed to nitpick, the poem transcription by ChatGPT was the one standout distinction — in any other case, they have been pretty evenly matched. I didn’t discover both lacking any crucial observations or misinterpreting something. Each instruments demonstrated sturdy comprehension and interpretation abilities.

ChatGPT transcribing my hand-written notes as a part of the picture evaluation job

So, on this spherical? It’s an in depth name. When you’re simply in search of quick, correct picture understanding, Perplexity completely holds its personal. However in the event you worth barely extra construction and that bonus degree of element like transcription, ChatGPT nudges forward.

Winner: ChatGPT

7. File evaluation

For this job, I gave each ChatGPT and Perplexity a heavy-hitter: Einstein’s 1905 paper “On the Electrodynamics of Transferring Our bodies” and requested them to summarize it in 5 bullet factors beneath 100 phrases.

ChatGPT’s response was polished, accessible, and grouped concepts like time dilation and size contraction effectively. It felt user-friendly and clear with out oversimplifying. It went a bit of above phrase depend however nonetheless not as a lot as Perplexity. 

ChatGPT’s response to the file evaluation job

Perplexity’s take leaned extra educational, utilizing phrases like “Lorentz transformations” and “mass-energy relationship” proper up entrance. It felt like one thing a researcher may write — exact, barely extra technical. It positively went a bit of over the phrase depend, too. 

Perplexity’s response to the file evaluation job

Winner? Discounting the phrase depend situation, I’d say it’s a tie. ChatGPT is nice for fast understanding. Perplexity feels extra formal. Each are wonderful at distilling advanced content material into digestible insights.

Winner: Cut up verdict. 

8. Information evaluation 

Information evaluation was subsequent. I supplied each with a CSV of U.S. ChatGPT search curiosity by subregion to see who might extract key insights.

And I’ve to say, Perplexity knocked it out of the park. It didn’t simply summarize the information; it broke it down with the form of element I didn’t get from ChatGPT, Gemini, and even DeepSeek. We’re speaking about statistical abstract with imply, median, and commonplace deviation, together with considerate observations beneath regional patterns.

The technocentric perception you see under? Tremendous helpful. It made me really feel like I used to be studying the evaluation of somebody who actually received the information, not simply glanced at it.

Perplexity’s response to the information evaluation immediate

ChatGPT did tremendous — clear, readable, and correct. If I wished a fast scan, positive, it delivered. But when I used to be prepping for a gathering or writing a report, I’d lean on Perplexity for the additional depth.

ChatGPTs response to the information evaluation immediate

This one wasn’t even shut. Clear win for Perplexity.

Winner: Gemini 

9. Actual-time internet search

Within the subsequent job, I used to be curious to see how effectively Perplexity and ChatGPT might sustain with the world. I requested them to search out and summarize the three most up-to-date AI information tales.

ChatGPT surprisingly rose to the occassion. It is actual time internet search pulled well timed headlines — White Home appointing Chief AI Officers, Meta launching Llama 4, and Microsoft showcasing an upgraded Copilot at its fiftieth anniversary occasion. Not solely have been the tales contemporary (all dated throughout the previous few days), however the summaries have been concise, structured, and filled with context.

ChatGPT’s response to the real-time internet search job

Perplexity, alternatively, took a extra analytical method. It pulled in deeper industry-wide tales just like the AI Index Report, Google’s Pixel 9 Gemini AI replace, and the way AI is reshaping MBA packages. These have been positively insightful and sourced from credible publications like Nature, however they weren’t as “news-now” as what I received from ChatGPT. Some headlines felt extra like current analysis findings or product commentary than hot-off-the-press updates.

ChatGPT’s response to the real-time internet search job

For me, ChatGPT got here out on prime for this one. 

Winner: ChatGPT

10. Deep analysis job

The ultimate job I designed was centered round what I imagine is a very pivotal functionality for AI chatbots: deep analysis. The promise of those instruments to deal with advanced analysis questions and effectively analyze huge quantities of data is extremely thrilling.

To check this immediately, I set each Perplexity and ChatGPT the problem of exploring a present and important space: the continuing traits in SaaS consolidation.

Perplexity responded shortly and packed its evaluation with up-to-date knowledge — 49 sources in whole. It nailed the numbers, cited current case research, and delivered a clear abstract with sturdy monetary context. The insights round tech hubs and valuation traits have been particularly sharp. That mentioned, it was extra of a straight knowledge drop — quick and correct.

Perplexity’s deep analysis capabilities

ChatGPT, in distinction, took its time. It requested me clarifying questions first, which made the expertise really feel extra collaborative. The ultimate report took about eight minutes, but it surely was well worth the wait. It pulled from 41 sources, included examples, and had a transparent strategic construction.

I did discover it leaned on older knowledge, which was a bit irritating since I’d requested for insights from the final 3–5 years. Nonetheless, the content material was wealthy and layered, with considerate takeaways for SaaS leaders and buyers.

ChatGPT’s Deep Analysis asking questions earlier than beginning

ChatGPT, alternatively, took a extra interactive method, asking me detailed questions on my most popular timeframe, geographic focus, and precedence areas earlier than continuing. In comparison with Gemini’s quicker turnaround, ChatGPT took longer to finish the duty (about eight minutes to generate the whole report).

Each instruments did very well, however in numerous methods. Perplexity is healthier for fast, data-heavy analysis. ChatGPT takes longer however offers you one thing nearer to an government briefing.

You could find each analysis reviews right here.

Winner: Cut up verdict

ChatGPT and Perplexity aren’t the one AI chatbots on the market. I’ve examined Claude, Microsoft Copilot,  and extra to see how they stack up in my greatest ChatGPT alternate options information. Test it out!