Llama 3.2 1B

1.2B parameters
reasoning
Parameters
1.2B
Context Length
128K
Architecture
llama
Release Date
2024

Available Quantizations

QuantizationFile SizeQuality
Q2_K0.5 GBLower
Q3_K_M0.6 GBLower
Q4_K_M0.7 GBGood
Q5_K_M0.8 GBGood
Q8_01.3 GBHigh

Can it run? (Q4_K_M quantization)

GeForce RTX 4090
Runs great
~970 tok/s • 24GB VRAM
GeForce RTX 4070
Runs great
~485 tok/s • 12GB VRAM
GeForce RTX 3060 12GB
Runs great
~346 tok/s • 12GB VRAM
Radeon RX 7900 XTX
Runs great
~924 tok/s • 24GB VRAM
Check with your specific GPU →