gpu4
.pics
show your rig · prove your tokens
Gallery
Leaderboard
VRAM fit
Submit your rig
About
Leaderboard
Generation speed ranked across accelerators — for a given model. Beautiful, not a spreadsheet.
Llama 3.2 1B Instruct
32
Meta Llama 3.1 8B Instruct
14
Qwen2.5 14B Instruct
12
DeepSeek R1 671B
2
Goliath 120B
2
Qwen3 235B
2
qwen3:14b
2
1
NVIDIA A100 80GB
80 GB ·community
12
tok/s
0.03 tok/s/W
2
Apple M2 Ultra
192 GB ·community
37
tok/s
—