GeForce RTX 5090, 32GB
1.8TB/s of memory bandwidth, and 31GB of GDDR7 memory left for a model once the display buffer and the CUDA context have taken theirs. It holds 14 of the 30 models we track.
This is a card, not a computer. Every price on this page is the card plus $1,000 for the minimum host machine around it, so the comparison with a whole machine is fair.
The card itself ranges from $1,999.99 for a Founders Edition to $7,070.58 for a board-partner card with a bigger cooler — the same 32GB either way. We quote the cheapest listing we can actually find.
Offers
checked 45 days ago- NVIDIAlist pricenot availablechecked 45 days ago$2,999
$1,999 for the card at NVIDIA, plus $1,000 for the host machine.
NVIDIA's published MSRP. We did not confirm stock at NVIDIA's own store, so we do not quote it as available.
- Best Buycheapest todayin stockchecked 45 days ago$2,999.99
$1,999.99 for the card at Best Buy, plus $1,000 for the host machine.
Founders Edition, at MSRP. The only card we found anywhere near $1,999.
- Best Buynot availablechecked 45 days ago$5,829.99
$4,829.99 for the card at Best Buy, plus $1,000 for the host machine.
ASUS ROG Astral. Stock unclear on the page; we did not assume it.
- Neweggin stockchecked 45 days ago$5,899.99
$4,899.99 for the card at Newegg, plus $1,000 for the host machine.
MSI Gaming Trio OC, sold by Newegg itself.
- Amazonnot availablechecked 45 days ago$6,768.88
$5,768.88 for the card at Amazon, plus $1,000 for the host machine.
Gigabyte WINDFORCE OC. Price shown, stock not stated on the page.
- Newegg Marketplacein stockchecked 45 days ago$7,130.80
$6,130.80 for the card at Newegg Marketplace, plus $1,000 for the host machine.
PNY Triple Fan, third-party seller.
- Newegg Marketplacein stockchecked 45 days ago$8,070.58
$7,070.58 for the card at Newegg Marketplace, plus $1,000 for the host machine.
MSI Ventus 3X OC, third-party seller, ships from Canada.
6 reseller listings for the card itself, from $1,999.99 at Best Buy to $7,070.58 at Newegg Marketplace — the same 32GB either way.
Timing
Wait. It has been $1,999 in the last 45 days, $0.99 below today.
- 28 Aug 2026$1,999.99
- 28 Aug 2026$4,829.99
- 28 Aug 2026$4,899.99
- 28 Aug 2026$5,768.88
- 28 Aug 2026$6,130.80
- 28 Aug 2026$7,070.58
Models it runs — 8K context
- Won't fit— headroom
- Won't fit— headroom
- Won't fit— headroom
- Won't fit— headroom
- Won't fit— headroom
- Won't fit— headroom
- Won't fit— headroom
- Won't fit— headroom
- Won't fit— headroom
- Won't fit— headroom
- Won't fit— headroom
- Won't fit— headroom
- Won't fit— headroom
- Won't fit— headroom
- Won't fit— headroom
- Won't fit— headroom
- Qwen3.5-35B-A3BQ6_KTight1GB headroomest. 364 tok/s
- Qwen3-32BQ6_KTight1GB headroomest. 44.4 tok/s
- Gemma-4-31BQ6_KTight3GB headroomest. 46.3 tok/s
- Qwen3-30B-A3BQ6_KTight4GB headroomest. 331 tok/s
- Gemma-3-27BQ6_KFits5GB headroomest. 52.6 tok/s
- Qwen3.6-27BQ8_0Tight0GB headroomest. 40.6 tok/s
- Gemma-4-26B-A4BQ8_0Tight2GB headroomest. 222 tok/s
- Tight3GB headroomest. 45.7 tok/s
- gpt-oss-20bQ8_0Fits7GB headroomest. 234 tok/s
- Phi-4Q8_0Fits13GB headroomest. 78.3 tok/s
- Gemma-4-12BBF16Fits5GB headroomest. 48.7 tok/s
- Fits12GB headroomest. 72.8 tok/s
- Gemma-3n-E4BBF16Fits15GB headroomest. 146 tok/s
- Qwen3-4BBF16Fits20GB headroomest. 146 tok/s
| Model | Best quant | Fit | Headroom | Est. speed |
|---|---|---|---|---|
| Llama-4-Behemoth | — | Won't fit | — | — |
| DeepSeek-V4-Pro | — | Won't fit | — | — |
| Kimi-K2-Instruct | — | Won't fit | — | — |
| DeepSeek-V3 | — | Won't fit | — | — |
| DeepSeek-R1 | — | Won't fit | — | — |
| Qwen3-Coder-480B-A35B-Instruct | — | Won't fit | — | — |
| Llama-4-Maverick-17B-128E | — | Won't fit | — | — |
| GLM-4.6 | — | Won't fit | — | — |
| DeepSeek-V4-Flash | — | Won't fit | — | — |
| Qwen3-235B-A22B | — | Won't fit | — | — |
| MiniMax-M2 | — | Won't fit | — | — |
| Mistral-Large-2411 | — | Won't fit | — | — |
| gpt-oss-120b | — | Won't fit | — | — |
| Llama-4-Scout-17B-16E | — | Won't fit | — | — |
| Qwen3-Coder-Next | — | Won't fit | — | — |
| Llama-3.3-70B | — | Won't fit | — | — |
| Qwen3.5-35B-A3B | Q6_K | Tight | 1GB | 364 tok/s |
| Qwen3-32B | Q6_K | Tight | 1GB | 44.4 tok/s |
| Gemma-4-31B | Q6_K | Tight | 3GB | 46.3 tok/s |
| Qwen3-30B-A3B | Q6_K | Tight | 4GB | 331 tok/s |
| Gemma-3-27B | Q6_K | Fits | 5GB | 52.6 tok/s |
| Qwen3.6-27B | Q8_0 | Tight | 0GB | 40.6 tok/s |
| Gemma-4-26B-A4B | Q8_0 | Tight | 2GB | 222 tok/s |
| Devstral-Small-2507 | Q8_0 | Tight | 3GB | 45.7 tok/s |
| gpt-oss-20b | Q8_0 | Fits | 7GB | 234 tok/s |
| Phi-4 | Q8_0 | Fits | 13GB | 78.3 tok/s |
| Gemma-4-12B | BF16 | Fits | 5GB | 48.7 tok/s |
| Qwen3-VL-8B-Instruct | BF16 | Fits | 12GB | 72.8 tok/s |
| Gemma-3n-E4B | BF16 | Fits | 15GB | 146 tok/s |
| Qwen3-4B | BF16 | Fits | 20GB | 146 tok/s |
Best quant is the highest-quality one that still loads with headroom. Every speed is an estimate; the formula is here.