HP has for
HP Employee Purchase Program Members (
register/login):
HP OMEN MAX 45L Gaming Desktop on sale from
$3956.62 when you follow the steps below and apply promo code
AWAKENING25 or
BYESUMMER25 at checkout.
Shipping is free.
Thanks to Community Member
Elon69 for sharing this deal.
Deal Instructions:
- Click here to register or login to your HP Employee Purchase Program (EPP) account.
- Visit the HP Omen 45L Max product page:
- Click 'Customize & Buy', then under configurations choose the following:
- Graphics Card: NVIDIA GeForce RTX 5090 +$1750
- Memory: Kingston FURY 32GB DDR5-6000 MT/s EXPO RGB Heatsink (2 x 16 GB) +$20
- Chassis and Power Supply: 1200 W Cybenetics power supply +$190
- Optional: Upgrade from Wi-Fi 6E + Bluetooth 5.3 to Wi-Fi 7 + Bluetooth 5.4 for +$5 extra
- Add to cart
- Apply 25% off promo code AWAKENING25 or BYESUMMER25 in cart
- Price will be as follows after automatic 5% off HP EPP discount and 25% off code are applied (without Wi-Fi upgrade):
- w/ Ryzen 7 9700X CPU (GT23-0000t) $3956.62
- w/ Ryzen 7 9800X3D CPU (GT23-0100m) $4119.74
Specs:
- CPU / Processor:
- AMD Ryzen 7 9700X 8-Core / 16-Thread Processor
- or AMD Ryzen 7 9800X3D 8-Core / 16-Thread Processor
- NVIDIA GeForce RTX 5090 32GB GDDR7 Graphics
- 32GB (2x16GB) DDR5-6000 Dual-Channel Memory / RAM
- 1TB PCIe Gen4 NVMe Solid State Drive Storage / SSD
- Wi-Fi 6E + Bluetooth 5.3
- AMD B850 Motherboard
- OMEN 45L Case & 1200W Cybenetics Platinum ATX 3.1 Fully Modular Power Supply
- Windows 11 Home
- Ports:
- Top:
- 1x USB-C (10Gbps, HP Sleep and Charge)
- 2x USB-A (5Gbps, HP Sleep and Charge)
- 1x Headphone / Microphone Combo
- Rear:
- 2x USB-C (10Gbps, one with DisplayPort 1.4)
- 2x USB-A (5Gbps)
- 4x USB-A 2.0
- 1x RJ45 Ethernet (2.5GbE)
- 1x Microphone
- 1x Line-in
- 1x Line-out
- GPU Outputs:
Warranty: 1-Year Limited Warranty
Top Comments
I've noticed the cheapest 5090s at Microcenter are now ~$4200. So with this deal, you're literally getting a whole PC for free w/5090 purchase.
I could only get it down to $4087 as only the 'awakenings' code worked for me. Still, pretty good!
The only thing proprietary about this entire AMD Omen is the front-panel I/O.
$6 part is all you need if you ever want to move the motherboard to a standard case.
https://www.amazon.com/dp/B0GMHRPQJX
Everything else is standard fit and ports etc. No funny connectors, no special PSU, no weird cables for fans etc. Whatever urban legends people have about HP proprietary parts.
GPU is usable in any other motherboard's PCI-E slot. Down to the side screw holes on 50xx long cards where it can be used with fancier brackets to secure the card down. Far better than a simple ant-sag stick under it. These screw holes is what HP used to secure the card down like a tank in the machine.
Lenovo has similar extra study GPU brackets. Nothing proprietary about the GPU card themselves that makes it so you can't take it and use it somewhere else. Clearly you have zero experience with these type of machines... lol
People who spread misinformation without bothering to check the facts are worse than a glitchy AI.
Join The Conversation
Share your experience with the Slickdeals community
Share information with the community. Please follow our Community Guidelines and be kind!
93 Comments
Sign up for a Slickdeals account to remove this ad.
LOL LOL
My CC bill on PC purchases will be 26K next month! Joy of 5090 hunting.
for local LLM vs.Online - the pay off isn't there for 99% of the users, you will always be better off in cost if you use online models and they are definite better than any local models. The only time local AI will come out ahead on cost will be video generation per my experience.
If you really want local AI, 27B is best. 35B AoE is next best and uses less VRAM.
Qwen 3.8 27B is the best sweet spot now overall, this will run just fine. Smaller models look pretty weak in comparison but you can run 35B MoE A3 off of 16GB RAM.
There is nothing like a 32GB (or penny's 96GB VRAM though).
You can also double up on 16GB cards for 32GB, I ran 5080+5070 TI also if you have slimmer 2 card or 2.5 card height GPUs and Mobo is enough and use a 1200w PSU.
Performance is ok and great to be able to do 32GB VRAM for around $2500 max cost across 2 cards. vs. $4000 5090
If you do dual 5070 TI, it could have been $1700 for 32GB total VRAM
I picked up a 3090 24GB 2 days ago, not sure how I want to deploy it yet. Maybe pair with a 16GB for a total of 40GB to parallel LLM and content creation. 3090 is much slower but but you can get to 48GB VRAM if you find 2 x 3090 24GB for $2000 total.
I did some basic testing on it, it feels like running LLM on a Macbook M5 Max, got more ram but it's also slower... not very satisfying.
It's a lot of pressure psychologically to pull the trigger because of increasing GPU prices. I think I'm going to wait it out still. The speed of local LLM progress and potential sweet spot in hardware requirement in a year's time is still up in the air. I might jump when it's more clear
I checked this idea with AI and it warns of fraud detection potentially declining payment or shipment.
(And if you're thinking to have shipping courier route to FedEx or UPS pick up location - AI says HP does not allow that).
If yall have better work around I'm all ears.
For Amazon purchases one can do pickup at an Amazon locker or counter in Delaware.
Sign up for a Slickdeals account to remove this ad.
This is exactly my use case. Chat / Text Completion isn't that great. You are better off with a $100 a month Claude subscription for chat and coding work.
But for video, there is no contest. Host LLMs can cost $30-40 a 2 minute video. I can build the same video for $2. And in some cases, image generation is a lot faster - 13 seconds with Z-Image vs Grok xAI round trip which can be 40 seconds.
I can do a 30 second 480p video with a lipsynced talking avatar with a RTX 5070ti. But with a 5090, I can do 720p then upscale.
I checked this idea with AI and it warns of fraud detection potentially declining payment or shipment.
(And if you're thinking to have shipping courier route to FedEx or UPS pick up location - AI says HP does not allow that).
If yall have better work around I'm all ears.
For Amazon purchases one can do pickup at an Amazon locker or counter in Delaware.
I am running a few boxes. Image generation on a 12GB 5070 & 16GB 5070ti - Z-Image, Flux both with nvfp4 models.
For video, I am running WAN 2.2 with InfiniteTalk for lipsync and LTX 2.5 which is highly distilled. LTX2.5 is super fast, I can get 5 second video in 70 seconds but it artifacts as hell so I go with WAN 2.2 for non human (3D Pixar, Stickman, Comic). LTX excels there. WAN 2.2 for anything human right now. I can't even run MiniMax H3 so the 5090 will allow me to try that.
I can produce an hour long Stickman style faceless youtube video with caption, narration, all for nothing but electricity cost. The kind you see on YT with biblical and historical story telling. Faceless Youtube.
Al I need to do is break one niche running 24/7. Produce three to 4 20 minutes videos or 30 1 minute TikTok/Instagram reels.
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!