expired
Lenovo Legion Pro 7i: 16" 2.5K 240Hz OLED, Intel Ultra 9 275HX, RTX 5090, 32GB DDR5, 1TB SSD $2999
$2,999
$3,999
25% off
B&H Photo Video
Missed This Deal? Get alerts when similar deals are posted.
B&H Photo VideoDeal Details
Includes 2 Free Items:
- Bitdefender Total Security (Download, 5 Devices, 6 Months)
- NVIDIA GeForce RTX 50 Series Pragmata Game Bundle
- 2.7 GHz Intel Core Ultra 9 24-Core
- 32GB of DDR5 RAM | 2TB M.2 SSD Storage
- 16" 2560 x 1600 OLED 240 Hz Display
- NVIDIA GeForce RTX 5090 (24GB GDDR7)
- Thunderbolt 4 | HDMI 2.1 | USB-C | USB-A
- 2.5 GbE | Wi-Fi 7 (802.11be) | BT 5.4
- 5MP Webcam with Privacy Shutter
- RGB Backlit Keyboard | Touchpad
- Windows 11 Home
Community Notes
About the Poster
Deal Details
Community Notes
About the Poster
Includes 2 Free Items:
- Bitdefender Total Security (Download, 5 Devices, 6 Months)
- NVIDIA GeForce RTX 50 Series Pragmata Game Bundle
- 2.7 GHz Intel Core Ultra 9 24-Core
- 32GB of DDR5 RAM | 2TB M.2 SSD Storage
- 16" 2560 x 1600 OLED 240 Hz Display
- NVIDIA GeForce RTX 5090 (24GB GDDR7)
- Thunderbolt 4 | HDMI 2.1 | USB-C | USB-A
- 2.5 GbE | Wi-Fi 7 (802.11be) | BT 5.4
- 5MP Webcam with Privacy Shutter
- RGB Backlit Keyboard | Touchpad
- Windows 11 Home
Community Voting
Deal
Score
+35
Good Deal
Bad Deal
Community Reviews
This Legion Pro 7i stands out as a high end gaming and AI laptop with a 16 inch 2.5K 240Hz OLED screen, RTX 5090 graphics, and enough memory and storage for heavy local workloads. The machine is described as unusually silent and cool for its specs, and the 5090 plus 64GB RAM setup is praised for local AI, video, and image generation. The main caveats are that the laptop 5090 is power limited and much slower than a desktop 5090 in some workflows, and Thunderbolt 4 instead of Thunderbolt 5 is a deal breaker for some. It looks strongest for portable performance when the workload fits inside 24GB of VRAM.
Common Questions (5)
AI-generated summary based on comments from the community. Updated ago
Join The Conversation
Share your experience with the Slickdeals community
Share information with the community. Please follow our Community Guidelines and be kind!
40 Comments
Sign up for a Slickdeals account to remove this ad.
395 AI Max 128GB both use SoC.
I would never get the AI Max, that's even slower than the Mac for LLM ( I think). and to me less functional as a daily driver. I don't like to use Windows as a UI and hates Linux UI even more. I live in the terminals a lot and cmd/powershell are crap and I don't want need to install another terminal emulator. Also more hassle to maintain a windows machine and software libraries for me - partly I just don't want to invest time to begin with. I have a large apple eco and I prefer it. I'm sure people can be very efficient on Windows/Linux also.
People liked AI Max 128GB cuz it was a lot cheaper than 128GB Mac and cheaper than 5090 machines. It was a way to run very large models which I don't find a use yet even though I bought M5 Max 128GB macbook. I think AI max is up to $3500 due to ram? it was $2000 when it got really popular upon release.
I don't find those 70B / 80B models more useful or smarter- this is evidenced by the newer Qwen 3.6 27B/35B models I think. I think the hype or "oh I can run very large models faded" because smaller models got so much smarter.
RTX in general. 3x peak memory bandwidth.
CUDA performance is always king, thousands of cores vs my 40 GPU cores. I know they're different but software heavily optimized by developers on CUDA first always and been done for years, then Metal MPS for Apple came but still far less support and hardware itself is just not fast enough for this kind of computation.
5090 desktop has 1600GB around here memory bandwidth speed. 3x this m5 max with 40 GPUs around 600 GB. And this affects the token counts the most
I just ran a dumb test just now... yes it is dumb
Qwen 35B MOE - just generating 1000 words of something.
Desktop 5090 32G VRAM - 225 tokens/second
M5 Max 40 GPUs (128GB) 118 tokens per second.
AI Max. no idea.
Images? lol. laughable.
I don't generate images on my Mac anymore but here are some examples that I'm trying to go by memory.
zImage (I think it was turbo idk...)
1024x1024 image on RTX Laptop 5090. less than 1 second per image.
1024x1024 image on M4 Macbook 20GPUs, I think it was 1 minute.. mayeb it was 2-33 minutes?
and RTX 2070 8GB (yes 2070 from 2019 era) did it in 10 seconds.
Generating MUSIC using that free stuff... HeartMula
- RTX laptop 5090 - 2 minutes?
- Macbook lack of support at the time - 30 minutes or 45 minutes. I can't remember and because it was going on the CPU, I can't really use my Mac cuz it was pegging the machine here and there.
Slow image generation I can deal with since it CPU are not pegged and I can use my Mac apps since they mostly use CPU cycles. But if the CPU is being used for AI work, I might as well go take a dump and come back. That's why I got RTX machines, I can't stand my Mac CPUs strained the point I can't use it for the various AI stuff I was playing with. And Windows is just terrible power management so I can't just buy a RTX laptop and go out with it. plus I prefer the apple eco system and macOS.
I just scanned this... but it seems to say AI Max on Qwen 3.5 (not 3.6) 35B was 27 tokens per second? That is utter garbage to me. One of a setting got up to 43 tokens/s it seems.
https://www.reddit.com/r/LocalLLa...?rdt
I'm still rocking my 4060 gaming laptop.
Everything is too expensive when it comes to gpu and ram.
Sign up for a Slickdeals account to remove this ad.
lol well, was seriously contemplating this deal but Memorial Day is only 2 weeks out.
Sign up for a Slickdeals account to remove this ad.
Join The Conversation
Share your experience with the Slickdeals community
Share information with the community. Please follow our Community Guidelines and be kind!