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
▣ AMD Instinct MI50 32GB
32 GB
gpt-oss 120B· Q8
58
tok/s
★ featured community
Apple M3 Max
128 GB
IBM Granite 4.1 8B· Q4_K_M
57.6
tok/s
593 · Strong ★ featured measured
Apple M2 Ultra
192 GB
Dolphin Mixtral 8x7B· Q5
37
tok/s
★ featured community
Apple M3 Ultra
512 GB
DeepSeek R1 671B· Q4
18
tok/s
★ featured community
▤ NVIDIA Tesla V100
32 GB
Qwen3 235B· Q4
15
tok/s
★ featured community
▤ NVIDIA L40S
44 GB
Llama 3.2 1B Instruct· Q4_K_M· 1.5B
355
tok/s
5,252 · Excellent LocalScore
▤ NVIDIA RTX 3090 Ti
24 GB
Llama 3.2 1B Instruct· Q4_K_M· 1.5B
354
tok/s
3,908 · Excellent LocalScore
▤ NVIDIA H100
80 GB
Llama 3.2 1B Instruct· Q4_K_M· 1.5B
335
tok/s
4,512 · Excellent LocalScore
▤ NVIDIA RTX 3090
24 GB
Llama 3.2 1B Instruct· Q4_K_M· 1.5B
330
tok/s
1.14 tok/s/W LocalScore
▤ NVIDIA RTX A6000
48 GB
Llama 3.2 1B Instruct· Q4_K_M· 1.5B
315
tok/s
3,438 · 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 3090
24 GB
Llama 3.2 1B Instruct· Q4_K_M· 1.5B
308
tok/s
3,294 · Excellent LocalScore
▤ NVIDIA Tesla V100
32 GB
Llama 3.2 1B Instruct· Q4_K_M· 1.5B
248
tok/s
2,545 · Excellent LocalScore
▤ NVIDIA Tesla V100
32 GB
Llama 3.2 1B Instruct· Q4_K_M· 1.5B
248
tok/s
2,545 · 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 RTX 4090
24 GB
Llama 3.2 1B Instruct· Q4_K_M· 1.5B
207
tok/s
2,559 · Excellent LocalScore
Apple M4 Max
128 GB
Llama 3.2 1B Instruct· Q4_K_M· 1.5B
184
tok/s
5.2 tok/s/W LocalScore
▤ NVIDIA RTX 5090
32 GB
Llama 3.2 1B Instruct· Q4_K_M· 1.5B
170
tok/s
3,245 · Excellent LocalScore
Apple M4 Max
128 GB
Llama 3.2 1B Instruct· Q4_K_M· 1.5B
156
tok/s
3.64 tok/s/W LocalScore
▤ NVIDIA RTX 6000 Ada Generation
48 GB
Llama 3.2 1B Instruct· Q4_K_M· 1.5B
131
tok/s
3,350 · 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 3090 Ti
24 GB
Meta Llama 3.1 8B Instruct· Q4_K_M
110
tok/s
1,113 · 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 RTX 3090
24 GB
Meta Llama 3.1 8B Instruct· Q4_K_M
103
tok/s
0.31 tok/s/W LocalScore
▤ NVIDIA H200
141 GB
Llama 3.2 1B Instruct· Q4_K_M· 1.5B
98.5
tok/s
2,539 · Excellent LocalScore
▤ NVIDIA RTX A6000
48 GB
Meta Llama 3.1 8B Instruct· Q4_K_M
90.5
tok/s
938 · Strong LocalScore
▤ NVIDIA RTX 4090
24 GB
Meta Llama 3.1 8B Instruct· Q4_K_M
90.3
tok/s
1,439 · 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 5090
32 GB
Meta Llama 3.1 8B Instruct· Q4_K_M
66.3
tok/s
1,193 · Excellent LocalScore
▤ NVIDIA RTX 3090 Ti
24 GB
Qwen2.5 14B Instruct· Q4_K_M· 14.8B
64.2
tok/s
627 · Strong LocalScore
▤ NVIDIA RTX 3090
24 GB
Qwen2.5 14B Instruct· Q4_K_M· 14.8B
59.9
tok/s
0.18 tok/s/W LocalScore
▤ NVIDIA RTX A6000
48 GB
Qwen2.5 14B Instruct· Q4_K_M· 14.8B
52
tok/s
521 · Strong LocalScore
Apple M4 Max
128 GB
Meta Llama 3.1 8B Instruct· Q4_K_M
51.6
tok/s
1.78 tok/s/W LocalScore
▤ NVIDIA RTX 6000 Ada Generation
48 GB
Meta Llama 3.1 8B Instruct· Q4_K_M
51.3
tok/s
1,038 · Excellent LocalScore
▤ NVIDIA RTX 4090
24 GB
Qwen2.5 14B Instruct· Q4_K_M· 14.8B
49.1
tok/s
766 · Strong LocalScore
▤ NVIDIA RTX 5090
32 GB
Qwen2.5 14B Instruct· Q4_K_M· 14.8B
45.5
tok/s
708 · Strong LocalScore
Apple M3 Max
128 GB
qwen3:14b
35.3
tok/s
submitted
▤ NVIDIA RTX 6000 Ada Generation
48 GB
Qwen2.5 14B Instruct· Q4_K_M· 14.8B
28.1
tok/s
560 · 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
▣ AMD EPYC 7352 24-Core Processor (znver2)
126 GB
Llama 3.2 1B Instruct· Q4_K_M· 1.5B
14.3
tok/s
88 · Modest LocalScore
▤ NVIDIA A100 80GB
80 GB
Goliath 120B· GGUF8-BIT
12
tok/s
0.03 tok/s/W community
▤ NVIDIA Tesla P40
24 GB
Qwen3 32B· Q4
5.9
tok/s
0.01 tok/s/W community