GeForce RTX 3080 10GB Price, VRAM & Local LLMs
Price alert
We'll email you when the eBay (US) price reaches your target.
NVIDIA's RTX 3080 10GB (2020) is a high-end gaming card that runs Gemma 4 12B at 4-bit at about 72 tokens/s, but it can't fit 27B-class models. Most cards on the market are used, so tell the 10GB, the 12GB and the modded 20GB versions apart.
eBay
GeForce RTX 3080 10GB price history
eBay
As an eBay Partner, VRAMGlass may be compensated if you buy through the eBay links on this page. The price you pay stays the same.
GeForce RTX 3080 10GB VRAM & full specs
- Memory10 GB · GDDR6X
- Bandwidth760 GB/s
- FP16 / BF16119.1 TFLOPS
- INT8238.1 TOPS
- Bus width
- 320-bit
- Architecture
- Ampere (GA102)
- Cores
- 8,704
- Tensor cores
- 272
- Ecosystem
- CUDA
- Interface
- PCIe 4.0 x16
- NVLink
- Not supported
- Power
- 320 W
- Power connector
- 12-pin（Founders Edition，附 2× 8-pin 转接线）
- Slots
- 2
- Length
- 285 mm
- Type
- Consumer
- Launch date
- Sep 17, 2020
- MSRP
- $699
Best GeForce RTX 3080 10GB deals
Powered by VRAMGlass real-time price tracking — handpicking the best-value hardware deals across the web
As an eBay Partner, VRAMGlass may be compensated if you buy through the eBay links on this page. The price you pay stays the same.
UsedMSI RTX 3080 Gaming Z Trio 10GB LHR - Working, Damaged 8-Pin Power Connector$399Buy on eBay
UsedHP GeForce RTX 3080 10GB OEM GPU - B1, B13D, B28 - Fan Defect, Only Displaypo...Top Rated$400Buy on eBay
UsedDELL NVIDIA GeForce RTX 3080 10GB GDDR6 Graphic Card$429Buy on eBay
UsedNVIDIA GeForce RTX 3080 GDDR6X 10GB Dedicated Graphics Card GPUTop Rated$430$622-31%Buy on eBay
UsedDell Alienware Nvidia RTX 3080 10GB Graphics Card GPU DesktopTop Rated$440Buy on eBay
UsedAORUS NVIDIA GeForce RTX 3080 10GB GV-N3080AORUS X-10GD HDMI/DP PCIe$450Buy on eBay
As an Amazon Associate I earn from qualifying purchases. The price you pay stays the same.
UsedHP NVIDIA GeForce RTX 3080 Graphics Card (Renewed)$500Buy on Amazon
UsedGaming OEM Nvidia GeForce RTX 3080 10GB GDDR6X 320-bit 19Gbps PCIe 4.0 x16 Graphics Video Card 0Y5K9F (Renewed)$530Buy on Amazon
UsedROG Strix NVIDIA GeForce RTX 3080 V2 OC Edition Gaming Graphics Card (PCIe 4.0, 10GB GDDR6X, LHR, HDMI 2.1, DisplayPort 1.4a, Axial-tech Fan Design, 2.9-Slot, Super Alloy Power II)(Renewed)$550Buy on Amazon
What LLMs can a single GeForce RTX 3080 10GB run?
A single RTX 3080 10GB is best with models up to 12B at 4-bit, where a 12B gets around 130K of context. 27B-class dense models don't fit even at 2-bit; 30B-class MoE models need their experts in system RAM, and Qwen3.6 35B A3B at 4-bit runs at about 51 tokens/s.
| Model | Size | Runs? | Speed t/s | Max context |
|---|---|---|---|---|
 Gemma 4 E4B8B | 3.8 GB | Runs | 124.2 | 128K |
 Qwen3.5 9B9.7B | 4.1 GB | Runs | 115.4 | 165K |
| Gemma 4 12B12B | 4.7 GB | Runs | 101.4 | 256K |
| Gemma 4 26B A4B25.2B (3.8B active) | 10.5 GB | Needs RAM offload | 54.7 | 256K |
| Qwen3.6 27B27.8B | 11.8 GB | Too big | — | — |
| Qwen3.8 27B27.8B | 9.8 GB | Too big | — | — |
 GLM 4.7 Flash30B (3B active) | 11.9 GB | Needs RAM offload | 50.6 | 146K |
| Gemma 4 31B30.7B | 11.8 GB | Too big | — | — |
| Qwen3.6 35B A3B36B (3B active) | 12.3 GB | Needs RAM offload | 76.6 | 256K |
| Model | Size | Runs? | Speed t/s | Max context |
|---|---|---|---|---|
| Gemma 4 E4B8B | 5 GB | Runs | 98.9 | 128K |
| Qwen3.5 9B9.7B | 5.7 GB | Runs | 87.6 | 118K |
| Gemma 4 12B12B | 7.1 GB | Runs | 71.5 | 132K |
| Gemma 4 26B A4B25.2B (3.8B active) | 16.9 GB | Needs RAM offload | 38 | 256K |
| Qwen3.6 27B27.8B | 16.8 GB | Too big | — | — |
| Qwen3.8 27B27.8B | 16.5 GB | Too big | — | — |
| GLM 4.7 Flash30B (3B active) | 18.3 GB | Needs RAM offload | 36.1 | 141K |
| Gemma 4 31B30.7B | 18.3 GB | Too big | — | — |
| Qwen3.6 35B A3B36B (3B active) | 22.1 GB | Needs RAM offload | 51.1 | 256K |
| Model | Size | Runs? | Speed t/s | Max context |
|---|---|---|---|---|
| Gemma 4 E4B8B | 8.2 GB | Runs | 64.1 | 85K |
| Qwen3.5 9B9.7B | 9.5 GB | Too big | — | — |
| Gemma 4 12B12B | 12.7 GB | Too big | — | — |
| Gemma 4 26B A4B25.2B (3.8B active) | 26.9 GB | Needs RAM offload | 25.7 | 256K |
| Qwen3.6 27B27.8B | 28.6 GB | Too big | — | — |
| Qwen3.8 27B27.8B | 29 GB | Too big | — | — |
| GLM 4.7 Flash30B (3B active) | 31.8 GB | Needs RAM offload | 22.6 | 131K |
| Gemma 4 31B30.7B | 32.6 GB | Too big | — | — |
| Qwen3.6 35B A3B36B (3B active) | 36.9 GB | Needs RAM offload | 33.9 | 256K |
RunsFits fully in VRAM; speed is estimated from VRAM bandwidth.
Needs RAM offloadMoE only: expert weights sit in system RAM, so speed is estimated from 70 GB/s RAM bandwidth.
Too bigWon’t fit on one card at this quantization, even with RAM offload.
Full calculation rules are on the Methodology page.
Who it suits
Good fit
- Fast generation with small models matters to you
Qwen3.5 9B at 4-bit runs at about 88 tokens/s with 118K of context
- You want small models at 8-bit
Gemma 4 E4B at 8-bit fits entirely on the card with 85K of context
- Upgrading an older motherboard
It's wired for all 16 PCIe 4.0 lanes, so it still gets 3.0 x16 on a PCIe 3.0 board
Not a good fit
- You want 27B-class models
Qwen3.8 27B at 2-bit needs 9.8 GB, and with context added it won't fit in 10GB of VRAM
- You want a 9B at 8-bit
Qwen3.5 9B at 8-bit needs 9.5 GB, and with context and runtime overhead added it won't fit in 10GB of VRAM
Before you buy
Don't mix up the three 3080s
The RTX 3080 12GB, released in 2022, has 2GB more VRAM, a 384-bit bus and 912 GB/s of bandwidth, so it runs models faster and the numbers on this page don't apply to it. A “3080 20G” in a listing is a modded card and isn't covered here either. If a listing just says “3080”, ask the seller for a GPU-Z screenshot and check for 10GB of VRAM and 8704 cores.
On a used card, check the memory temperature first
The 3080's GDDR6X memory runs hot, and on a card that's been in use for a few years the thermal pads may have degraded. Before buying, ask the seller to run a load for a while and watch Memory Junction Temperature (the VRAM temperature) in GPU-Z; be careful if it won't come down.
At 320W, the card needs a solid power supply
NVIDIA recommends a 750W power supply for the system. The Founders Edition is dual-slot and 285 mm long, powered through a 12-pin connector, with an adapter for two 8-pin cables in the box. Most partner cards use two or three 8-pin connectors directly and are longer, so measure your case before installing one.
Compared with similar GPUs
GeForce RTX 3080 10GB vs GeForce RTX 4070, GeForce RTX 3090, and GeForce RTX 2080 Ti 22GB: price, VRAM, bandwidth, compute and local LLM speed compared.
| Spec | GeForce RTX 3080 10GB | GeForce RTX 4070 | GeForce RTX 3090 | GeForce RTX 2080 Ti 22GB |
|---|---|---|---|---|
| Memory | 10 GB | 12 GB20% more | 24 GB140% more | 22 GB120% more |
| Bandwidth | 760 GB/s | 504 GB/s34% lower | 936 GB/s23% higher | 616 GB/s19% lower |
| BF16 / FP16 compute | 119 TFLOPS | 117 TFLOPSSame | 142 TFLOPS20% higher | 108 TFLOPS10% lower |
| INT8 compute | 238 TOPS | 234 TOPSSame | 285 TOPS20% higher | 215 TOPS10% lower |
| Qwen3.8 27B speed (2-bit) | Not supported | 36.4 t/s | 64.8 t/s | 44 t/s |
| Largest dense model (4-bit) | 13B | 16B23% more | 37B185% more | 34B162% more |
| Power | 320 W | 200 W38% lower | 350 W9% higher | 250 W22% lower |
| GPU interconnect | Not supported | Not supported | NVLink 112.5 GB/s | NVLink 100 GB/s |
| Interface | PCIe 4.0 x16 | PCIe 4.0 x16 | PCIe 4.0 x16 | PCIe 3.0 x16 |
| Released | 2020-09 | 2023-04 | 2020-09 | 2018-09 |
| eBay median |
As an eBay Partner, VRAMGlass may be compensated if you buy through the eBay links on this page. The price you pay stays the same.
Amazon.jp
Xianyu