llamaperf

Instinct MI250X 128GB vs Instinct MI300X 192GB

For running local LLMs · 1 report across 1 model

Side A
Instinct MI250X 128GB
Vendor
amd
VRAM
128GB
Memory
Discrete
Side B
Instinct MI300X 192GB
Vendor
amd
VRAM
192GB
Memory
Discrete

Highest reported generation speed by model family

These maxima come from different community setups. Model size, quantization, GPU count, offloading and concurrency may differ. They do not establish which GPU is faster for the same workload. Check the comparison conditions →

ModelInstinct MI250X 128GBInstinct MI300X 192GB
Qwen3.8up to 2400Bn=1

More comparisons