Wanted to share a tip, and confirm accuracy of my research while hunting for past unicorn deals of selected models I am interested in (Hint: Brokers, if you can beat these deals, hmu)
So far, the numbers look trustworthy based on the manual research I’ve been doing using the search bar since the last 2 weeks.
Examples of best deals ChatGPT could find:
(Here’s the ChatGPT Deep research link in case you want to check
You’re basically just looking at the edge cases of edge cases over all time - when we’re in a market that changes monthly, if not daily. You’re not going to have much success at all like this, and I see plenty of inaccuracies already from what it spit out (12/10k term on porsche doesn’t exist, etc.)
Correct. Best case scenarios for sure, but helps with a general baseline looking at Top N deals.
And you’re right, a lot of this relies on user generated posts that the LLM cant parse that well. For the Porsche, it comes from a reddit comment , Where the person later clarifies they paid $3k down.
3K down makes it $1100/month effective, which ChatGPT also didn’t catch. As good as ChatGPT is, it’s no replacement for comprehension and quality research.
Now I say this while using ChatGPT for a ton of stuff. This just isn’t what it’s good for.
@HersheySweet Ofcourse ofcourse. This is one edge case miss. but looking at the EQE EQS deals, seems pretty much on point. Nearly spent 2 weeks trying to understand the leasing landscape for MB, and pretty much came to the same conclusion as what the LLM spat out. But absolutely, nothing beats quality research (I just wish I had the time to do it for every car I cared about)
@ecooke2 : I pay $200 for the pro, and used deep research here. I think you get 3-4 per month with the $20 plan as well, but I am not sure.
Prompt :
I am looking for the best lease deal on Mercedes Benz EQB, EQE, EQS, Porsche Macan, and hyundai ioniq 5. Please go through all the pages in leasehackr forum under all sub topics like marketplace, signed deals and private transfers.
I want you to create a result that has top past/future 30 cheapest deals for all the makes and models I asked above. I want you to provide me links to the results as well as all information that will be valuable for me.
Also, check other forums like swapalease to find something competitive. Additionally, based on the results, recommend me the best 5 deals I could go with that’s currently available.
If I were to do this again, would’ve asked for a much structured response ( and seperate past vs current active leases )
How are y’all able to find the best available past and future deals by browsing marketplace for 5-10 minutes. Send me your tips. It has definitely taken me way way longer to understand the landscape of what price range I should be looking at
Keep in mind there could also be confirmation bias here. Folks who’re mentioning that “this should take 5 mins” might also be folks who spend a lot of time on this forum already, and the numbers might be second nature for them.
Or yeah, could be my skill gap. But overall I’m happy with this chatgpt research that took me 30 seconds of promoting
I think we just found an even worse answer than mentioning leasehackr to when dealers asked where you got your numbers… “ChatGPT said someone got this deal 5 years ago, so are you going to match it or am I going to have to take my business elsewhere?”
IDK about anyone else but when looking at marketplace I’m looking for today’s deals. Not past or future.
Is it really that hard to compare the deals for EQB and Ioniq 5 for example? Most listings are $ DAS and $ monthly. Total cost including broker fee for the same term and miles should be a pretty easy numerical calculation.
What prompt are you using to have ChatGPT normalize the data for differences in individual incentive qualifications/programs/tax rates/regional fees/etc?
I can’t even get it to accurately calculate a lease when all the proper inputs are given.
I don’t think it’s a skill gap. You seem intelligent to perform the numerical analysis I mentioned above by yourself or by prompting AI.
I think what will get you a lot closer to having a car in your driveway is
being more decisive in your choices (EQB vs EQS vs Macan are extremely divergent) and narrowing your focus accordingly
Understanding that payments fluctuate based on fluctuating RV/MF as John mentioned and eliminate any energy spent on creating what you presumably thought was a mind-blowing analysis of past deals.
Acknowledge that an AI that couldn’t understand how much the DAS affected your analysis or couldn’t convert a one-pay into an effective monthly is really useless for your goals right now
