Qwen 2.5 7B

7.6B parameters
reasoningcodingmultilingual
Parameters
7.6B
Context Length
128K
Architecture
qwen2
Release Date
2024

Available Quantizations

QuantizationFile SizeQuality
Q2_K3.0 GBLower
Q3_K_M3.8 GBLower
Q4_K_M4.7 GBGood
Q5_K_M5.4 GBGood

Can it run? (Q4_K_M quantization)

GeForce RTX 4090
Runs great
~153 tok/s • 24GB VRAM
GeForce RTX 4070
Runs great
~77 tok/s • 12GB VRAM
GeForce RTX 3060 12GB
Runs great
~55 tok/s • 12GB VRAM
Radeon RX 7900 XTX
Runs great
~146 tok/s • 24GB VRAM
Check with your specific GPU →