Datasheet for local AI49 models98 GPUsData read 2026-09-25
No adsNo tracking
Check my GPU →

RTX 3080 10 GB vs RTX 3060 12 GB for local AI

The RTX 3060 12 GB has 2 GB more memory, and that decides it for local AI: it runs a better file (or runs at all) for 32 of 49 models, and 17 run at 8-bit or better against 12.

RTX 3080 10 GB
VRAM
10 GB
Memory
GDDR6X
Bandwidth
760 GB/s
Architecture
Ampere
FP8 compute
No
Launch price
$699

All models on it →

VS
RTX 3060 12 GB
VRAM
12 GB
Memory
GDDR6
Bandwidth
360 GB/s
Architecture
Ampere
FP8 compute
No
Launch price
$329

All models on it →

01Model by model

ModelRTX 3080 10 GBRTX 3060 12 GB
image models
FLUX.1 devRuns Q4_K_SRuns Q5_K_S
FLUX.1 schnellRuns Q4_K_SRuns Q5_K_S
FLUX.1 KontextRuns Q4_K_MRuns Q5_K_M
FLUX.1 KreaRuns Q4_K_MRuns Q5_K_M
FLUX.1 FillRuns Q4_K_SRuns Q5_K_S
FLUX.2 devNot practical Q2_KOffload only Q4_K_M
FLUX.2 klein 9BRuns Q5_K_MRuns well FP8
FLUX.2 klein 4BRuns well 16-bitRuns well 16-bit
Krea 2Tight Q3_K_MRuns Q5_K_M
Qwen-ImageTight Q2_KTight Q2_K
Qwen-Image-EditOffload only Q2_KTight Q2_K
Qwen-Image 2.1Runs well Q8_0Runs well Q8_0
Z-Image TurboRuns well Q8_0Runs well Q8_0
Z-ImageRuns well Q8_0Runs well Q8_0
Ideogram 4Offload only Q4_1Offload only Q4_1
Boogu-ImageOffload only Q5_1Runs Q5_1
ERNIE-ImageRuns Q6_KRuns well Q8_0
HiDream-O1Offload only FP8Runs well FP8
Mage-FlowRuns well INT8Runs well 16-bit
Lumina 2.0Runs well 16-bitRuns well 16-bit
HiDream-I1 FullTight Q2_KTight Q3_K_M
HiDream-I1Tight Q2_KTight Q3_K_M
SD 3.5 LargeRuns Q5_1Runs well Q8_0
SD 3.5 MediumRuns well 16-bitRuns well 16-bit
Chroma1-HDRuns Q6_KRuns well FP8
SDXLRuns well 16-bitRuns well 16-bit
Illustrious / PonyRuns well 16-bitRuns well 16-bit
SD 1.5Runs well 16-bitRuns well 16-bit
HunyuanImage 2.1Offload only Q2_KTight Q2_K
video models
Wan 2.1 14BOffload only Q3_K_MTight Q3_K_M
Wan 2.1 1.3BRuns well 16-bitRuns well 16-bit
Wan 2.1 I2V 480POffload only Q4_K_MOffload only Q3_K_M
Wan 2.1 I2V 720POffload only Q4_K_MOffload only Q4_K_M
Wan VACE 14BOffload only Q4_K_MOffload only Q3_K_S
Wan 2.2 T2VTight Q2_KTight Q3_K_M
Wan 2.2 I2VTight Q2_KTight Q3_K_M
Wan 2.2 5BRuns well Q8_0Runs well Q8_0
Wan 2.2 AnimateOffload only Q2_KTight Q2_K
Wan Animate 2Offload only Q2_KTight Q2_K
Wan 2.2 S2VOffload only Q4_K_MOffload only Q4_K_M
SCAIL-2Offload only Q4_K_MOffload only Q2_K
HunyuanVideo 13BOffload only Q3_K_MTight Q3_K_M
LTX-Video 13BTight Q2_KTight Q3_K_M
HunyuanVideo 1.5Runs Q4_K_MRuns Q6_K
LTX-2Offload only Q4_K_MOffload only Q2_K
LTX-2.3Offload only Q4_K_MOffload only Q2_K
LTX-2.5Offload only Q4_K_MOffload only Q2_K
MiniMax H3Offload only Q4_K_MOffload only Q4_K_M
MiniMax H3 PrunedOffload only Q4_K_MOffload only Q4_K_M

Highlighted: the GPU that runs a better file for that model. Launch prices are the maker’s original list prices, not today’s street prices. Calculated from real file sizes; how the numbers work.

02Score

12run well on RTX 3080 10 GB
17run well on RTX 3060 12 GB
0better on RTX 3080 10 GB
32better on RTX 3060 12 GB