gpu4.pics
data via LocalScore
NVIDIA H200
141 GB VRAM NVIDIA 700 W TDP
running Llama 3.2 1B Instruct· Q4_K_M· 1.5B
98.5
tokens / sec
2,539 · Excellent #40 of 768
Generation
98.5tok/s
higher is better
Prompt / prefill
14,696tok/s
higher is better
Time to first token
89 ms
lower is better
Composite score
2,539
Model size
1.5B
Runtime
gpu4.pics/r/ls-nvidia-h200-llama-32-1b-instruct-q4-k-m

All benchmarks on this rig

ModelQuantGenPromptTTFTPowerPerf/W$/MtokScore
Llama 3.2 1B Instruct1.5BQ4_K_M98.514,69689 ms2,539

How it stacks up — Llama 3.2 1B Instruct

NVIDIA H200
98.5
NVIDIA L40S
355
NVIDIA RTX 3090 Ti
354
NVIDIA H100
335
NVIDIA RTX 3090
330
NVIDIA RTX A6000
315
NVIDIA A100 80GB
308