← Back to GPUs
NVIDIA · RTX 30
NVIDIA GeForce RTX 3060 12GB
$230$329 MSRP
The NVIDIA GeForce RTX 3060 12GB has become a legend in the budget AI community. Despite its modest gaming performance, the 12GB of VRAM with full CUDA support makes it the cheapest entry point for running local LLMs. It handles 7B-8B models at Q4-Q8 and runs Stable Diffusion 1.5. Available used for around -230, it is the go-to recommendation for AI beginners on a tight budget.
Best ForCheapest possible entry into local AI with CUDA
VerdictThe people's AI GPU — 12GB CUDA for used. Start here if budget is everything.
AI
4/10
Gaming
5/10
Specifications
VRAM12GB GDDR6
Memory Bandwidth360 GB/s
CUDA Cores3,584
Boost Clock1777 MHz
TDP170W
Power Connector1x 8-pin
Length242mm
Form FactorDual Slot
Release Year2021
AI Capabilities
Entry Level12GB VRAM
Limited to small models with heavy quantization. Fine for experimenting.
Can run (Q4 quantized)
Llama 3.1 8BQwen 2.5 14BMistral 7BFLUX.1 DevStable Diffusion XLStable Diffusion 3.5 LargeCogVideoX-5BMochi 1LTX VideoStable Video DiffusionWan Video 14BAlphaFold 2ESMFold (ESM-2 15B)ESM-2 3BscGPTRFdiffusionFine-tune Llama 8BTrain SDXL LoRA
Tight fit (may need CPU offload)
HunyuanVideo (14GB Q4)Codestral 22B (13GB Q4)Train FLUX LoRA (16GB Q4)
Recommended system RAM for AI: 24GB+ (2x GPU VRAM for model overflow)
Performance Estimates
Estimated tokens/sec for LLM inference based on 360 GB/s memory bandwidth — not hardware benchmarks. Methodology · What is Q4/Q8?
Llama 3.1 8B8B
Q8~23-29 tok/sUsableQwen 2.5 14B14B
Q4~22-27 tok/sUsableMistral 7B7B
Q8~27-33 tok/sUsableCodestral 22B22B
Offload~1-3 tok/sVery slowPros
- +12GB VRAM at budget price
- +Popular for entry-level AI
- +Very affordable used
Cons
- -Slow for modern gaming
- -Old architecture
- -Low bandwidth
budgetai
Where to Buy
Will It Run?
Llama 3.1 8B8B
Q8Qwen 2.5 14B14B
Q4Mistral 7B7B
Q8FLUX.1 Dev12B
Q4Stable Diffusion XL6.6B
Q8Stable Diffusion 3.5 Large8B
Q8HunyuanVideo13B
OffloadCogVideoX-5B5B
Q8