Can an AI build a gaming PC? MeinMMO editor Benedikt Schlotmann wanted to know if a chatbot could provide him with the right configuration. It didn’t work out at first, as none of the chatbots, ChatGPT, Perplexity, and Gemini, managed it flawlessly. Nevertheless, all provided sensible information.
For many PC builders, especially beginners, choosing the right components poses a significant hurdle. There is a lot of selection, and the high prices have not made the situation any better, but rather even more challenging.
I thought to myself: Maybe a chatbot can build the perfect gaming PC for me without needing much experience. Linus Tech Tips tried this a few years ago as well. Overall, I tested three bots: ChatGPT (free), Perplexity (free), and Google’s Gemini (free, Flash). None managed it right off the bat.
You can read about my adventurous journey with the chatbots through the world of gaming PCs in this article. I asked all three bots the following question:
Hello! I would like to build a mid-range gaming PC. My budget is 1,500 euros, the PC should handle WQHD resolution with stable 100 FPS, lowering graphics settings is a valid option. What hardware would you recommend to me in July 2026?
As far as possible, I also asked all bots identical questions when I had follow-ups to make the results comparable. How well that worked out, you can read in the following.
ChatGPT: Good configuration but stubborn about incorrect prices
Basically, the configuration presented by ChatGPT is quite reasonable. For the CPU, ChatGPT chose a mid-range CPU with an attractive price-performance ratio, and for the GPU, the RX 9070 XT, which I currently also consider the best mid-range option.
ChatGPT’s 1st configuration (1,500 euros)
| Component | Recommendation | Street price (approx.) |
|---|---|---|
| CPU | AMD Ryzen 7 9700X | 270–310 € |
| CPU cooler | Thermalright Phantom Spirit 120 SE or Peerless Assassin 120 SE | 40–50 € |
| Mainboard | B850 ATX (MSI Tomahawk / Gigabyte Eagle / ASUS TUF) | 170–210 € |
| RAM | 32 GB (2×16 GB) DDR5-6000 CL30 | 90–110 € |
| Graphics card | AMD Radeon RX 9070 XT 16 GB | 650–720 € |
| SSD | 2 TB PCIe-4.0 NVMe (WD SN850X, Lexar NM790, Kingston KC3000, etc.) | 100–130 € |
| Power supply | 750 W, 80+ Gold (be quiet!, Corsair RMx, Seasonic Focus) | 100–120 € |
| Case | Fractal Pop Air, Montech Air 903, Corsair 4000D Airflow, Phanteks XT Pro Ultra | 80–110 € |
The problem, however, is the price indications: I can currently find nowhere 32 GB DDR5 for 110 euros and a 2 TB M.2 SSD either.
Upon inquiry, the chatbot then realizes that it miscalculated and offers me updated prices: These are only 30 euros higher than before. Additionally, ChatGPT still offers the same configuration again but points out that I could also opt for the version without XT.
Only after I consciously tell the AI the current prices does ChatGPT recalculate the build with the updated prices and can recommend me an adjusted system. Interestingly, ChatGPT is the only bot that recommends a system with only 16 GB RAM, arguing that with performance:
This is one of the rare times when I would consciously only install 16 GB of RAM. Normally, I wouldn’t recommend that in 2026 – but currently, the upgrade to 32 GB costs around 200 euros more, while in games it often only provides a few percentage points of more performance.
Moreover, ChatGPT advises me that I can later upgrade if prices fall again.
Gemini: Incorrect prices, sold out graphics card, but then it works
Gemini also offers me a reasonable configuration right off the bat. Interestingly, Gemini is inherently more economical and only offers me the standard 9070 without XT and also the more affordable AMD Ryzen 5 9600X, which is usually quite sufficient for gaming.
Gemini’s 1st configuration (1,500 euros)
| Component | Recommendation | Approximate Price |
| Processor (CPU) | AMD Ryzen 5 9600X (6 cores / 12 threads) | ~170 € |
| Graphics card (GPU) | AMD Radeon RX 9070 (16 GB GDDR6) | ~580 € |
| CPU cooler | Arctic Freezer 36 | ~25 € |
| Mainboard | MSI B650 Gaming Plus WIFI (or Gigabyte B650 Eagle AX) | ~140 € |
| Memory (RAM) | 32 GB DDR5-6000 CL30 (e.g. G.Skill Flare X5) | ~100 € |
| Hard drive (SSD) | 2 TB M.2 NVMe SSD (e.g. WD Blue SN580 or Lexar NM790) | ~110 € |
| Power supply (PSU) | 750W be quiet! Pure Power 12 M (ATX 3.0 / fully modular) | ~100 € |
| Case (Case) | Endorfy Arx 500 Airflow (or Montech Air 903 Base) | ~75 € |
| Total price | Self-build price | ~1,300 € |
However, Google’s AI makes the same mistake as ChatGPT and provides completely utopian prices. 100 euros for RAM and SSD are simply not available anywhere today.
Upon inquiry, Gemini realizes it has made an error and correctly updates the prices for RAM but also replaces the RX 9070 with an RX 7800 XT (16 GB). However, that’s a graphics card you can’t buy anymore today. And this has to be known; otherwise, you end up buying an outdated graphics card or simply find that you can’t buy it anywhere anymore.
In a third round, Gemini replaces the 7800 XT with an RX 9060 XT and reduces the 2 TB to 1 TB to stay around 110 euros.
The final build from Gemini then amounts to 1,340 euros and is indeed sensibly applicable: Specifically, Google’s chatbot offers me an AMD Ryzen 5 9600X with an RX 9060 XT, with 1 TB SSD. But without further inquiries, I wouldn’t have gotten that far.
By the way, this is my inquiry about the prices. I also asked all chatbots this way.
Thank you for the recommendations. However, the prices you provided are incorrect: I cannot currently get 32 GB RAM for 110 euros or the 2 TB NVMe SSD either. Can you check the data again with current market prices and tell me which hardware you can recommend so that I stay under 1,500 euros?
Perplexity: Two builds, incorrect prices, and a good result
Perplexity not only offers me one build but two: a system with an AMD card and an Nvidia setup. Both proposed builds are absolutely okay in terms of hardware; I personally like the one from Perplexity the most because it truly is the best mid-range build right now since the bot opted for an X3D processor. This configuration would also be my current recommendation for 2026 if someone wants to buy a gaming PC.
As an alternative to the RX 9070 XT from AMD, Perplexity also mentions either the RTX 5070 Ti or the RTX 5070 but points out that I might have to pay extra here.
Perplexity’s 1st configuration (1,500 euros)
| Part | Recommendation | Target price |
|---|---|---|
| CPU | AMD Ryzen 7 7800X3D | 279 to 324 euros |
| CPU cooler | Solid tower cooler, e.g. Thermalright Peerless Assassin class | 35 to 50 euros |
| Mainboard | B650 or B650E with Wi-Fi | 130 to 180 euros |
| RAM | 32 GB DDR5-6000 CL30/32 | 95 to 125 euros |
| SSD | 2 TB NVMe PCIe 4.0 | 100 to 140 euros |
| Graphics card | Radeon RX 9070 XT 16 GB | 645 to 690 euros |
| Power supply | 750 W, 80+ Gold | 90 to 130 euros |
| Case | Airflow Mid-Tower | 70 to 120 euros |
However, the prices are again an issue: because they are not correct either at Perplexity. Upon inquiry, Perplexity then immediately offers me an updated setup with new prices and aligns with realistic market prices.
Interestingly, the chatbot first offers me the old build with updated prices, explains that this is too expensive, and then offers me a new, updated setup with appropriate prices. This way, I can see why the old build is too expensive and where I could save.
In the end, the bot finally offers me an AMD Ryzen 5 7600 and an RX 9060 XT (16 GB) and optionally a reduction of the SSD to 1 TB (instead of 2 TB). The price then fits exactly.
Conclusion: All chatbots provide good information but struggle with the current price situation
How do the bots perform? All three bots immediately suggest a reasonable configuration that can generally be recommended. Perplexity additionally offers an Nvidia build alongside AMD, which is a nice plus point.
For all three: They all logically justify why they chose certain hardware. Therefore, the recommendations are understandable, and all three bots strongly advise to choose a graphics card with 12 or even better 16 GB VRAM.
However, it is generally the case that I have to ask follow-up questions with all three bots, which at least requires me to have the prices already in my head. A beginner who just wants a list of parts to order or input into a configurator may quickly encounter a problem because the prices do not match.
If the prices weren’t so volatile and extremely high right now, the answers would indeed be quite useful without follow-up questions.
Which chatbot wins? Back to my inquiry: I wanted to have a gaming system assembled for 1,500 euros in July 2026. Who wins in comparison?
- If I look at the path to the goal and the results, then Perplexity wins. The final build is recommendable, the prices are correct, and I could choose the hardware or order it as a pre-built.
- ChatGPT stubbornly sticks to incorrect prices; only when I mention the specific prices does the bot’s result become usable. ChatGPT is the only bot that questions a new purchase and advises me to opt for used hardware or to use my old PC for another 6 to 12 months.
- Gemini provides me with incorrect information in an intermediate step, which probably wouldn’t have been noticed by everyone. In the worst case, Gemini would have persuaded me to buy an outdated and overpriced graphics card. If I even could have obtained it. However, Gemini is the only one that prefers to calculate more economically rather than maxing out the budget.
Gemini and ChatGPT are more concrete overall regarding the individual components and directly provide products. Perplexity remains a bit more general.
What is actually better: building a PC yourself or buying a pre-built one? Among gamers, this is a frequent discussion that usually ends up with many proud of building. MeinMMO editor Benedict Grothaus sees it differently – and has had disputes in the editorial office because of this: I used to build PCs myself – today I would never do that again
Your opinion is important to us!
Do you like the article? Then let us know!