gpu4.pics show your rig · prove your tokens

Show your rig. Prove your tokens.

A beautiful, open gallery of homelab & local-AI rigs — real tokens/sec, perf-per-watt, and cost, on the hardware people actually run. 70 rigs and counting.

Submit your rig → Will my GPU run it?
FastestMost efficientBest scoreNewest
All brands AMDAppleNVIDIAUnknown
Any VRAM≥16 GB≥24 GB≥48 GB≥80 GB
▤ NVIDIA H100
80 GB
Llama 3.2 1B Instruct· Q4_K_M· 1.5B
335
tok/s
4,512 · Excellent LocalScore
▤ NVIDIA A100 80GB
80 GB
Llama 3.2 1B Instruct· Q4_K_M· 1.5B
308
tok/s
3,892 · Excellent LocalScore
▤ NVIDIA RTX PRO 6000 Blackwell Workstation Edition
95 GB
Llama 3.2 1B Instruct· Q4_K_M· 1.5B
215
tok/s
3,948 · Excellent LocalScore
▤ NVIDIA H100
80 GB
Meta Llama 3.1 8B Instruct· Q4_K_M
120
tok/s
1,382 · Excellent LocalScore
▤ NVIDIA A100 80GB
80 GB
Meta Llama 3.1 8B Instruct· Q4_K_M
110
tok/s
1,054 · Excellent LocalScore
▤ NVIDIA RTX PRO 6000 Blackwell Server Edition
95 GB
Llama 3.2 1B Instruct· Q4_K_M· 1.5B
106
tok/s
2,962 · Excellent LocalScore
▤ NVIDIA RTX PRO 6000 Blackwell Workstation Edition
95 GB
Meta Llama 3.1 8B Instruct· Q4_K_M
103
tok/s
1,616 · Excellent LocalScore
▤ NVIDIA H200
141 GB
Llama 3.2 1B Instruct· Q4_K_M· 1.5B
98.5
tok/s
2,539 · Excellent LocalScore
▤ NVIDIA H100
80 GB
Qwen2.5 14B Instruct· Q4_K_M· 14.8B
72.2
tok/s
802 · Strong LocalScore
▤ NVIDIA RTX PRO 6000 Blackwell Workstation Edition
95 GB
Qwen2.5 14B Instruct· Q4_K_M· 14.8B
70.8
tok/s
1,080 · Excellent LocalScore
▤ NVIDIA A100 80GB
80 GB
Qwen2.5 14B Instruct· Q4_K_M· 14.8B
66.3
tok/s
614 · Strong LocalScore
▤ NVIDIA RTX PRO 6000 Blackwell Server Edition
95 GB
Qwen2.5 14B Instruct· Q4_K_M· 14.8B
17.9
tok/s
438 · Passable LocalScore
▤ NVIDIA A100 80GB
80 GB
Goliath 120B· GGUF8-BIT
12
tok/s
0.03 tok/s/W community