BBrainOutput

Best hardware to run Qwen2.5 72B

The minimum self-hostable option and a comfortable recommended build for Qwen2.5 72B, plus the top devices that run it well.

Minimum

Apple Mac mini (M4 Pro)

Fits at Q4_K_M (~44GB) but limited bandwidth makes token generation slow for a 72B model.

Recommended

Supermicro 8x H100 SuperServer

Fits at FP16 (~145GB) with ~418.2GB headroom — about 3 concurrent instances.

Top devices that run it well

  • Supermicro 8x H100 SuperServer
    Supermicro · AI Servers

    Fits at FP16 (~145GB) with ~418.2GB headroom — about 3 concurrent instances.

    FP16 · ~145GBRuns well
  • Dell PowerEdge XE9680
    Dell · AI Servers

    Fits at FP16 (~145GB) with ~418.2GB headroom — about 3 concurrent instances.

    FP16 · ~145GBRuns well
  • AMD Instinct MI300X
    AMD · Datacenter GPUs

    Fits at FP16 (~145GB) with ~24GB headroom — about 1 concurrent instance.

    FP16 · ~145GBRuns well
  • NVIDIA H200 (141GB)
    NVIDIA · Datacenter GPUs

    Fits at Q8_0 (~78GB) with ~46.1GB headroom — about 1 concurrent instance.

    Q8_0 · ~78GBRuns well
  • Cloud H200 141GB (profile)
    Cloud · Cloud GPU Profiles

    Fits at Q8_0 (~78GB) with ~46.1GB headroom — about 1 concurrent instance.

    Q8_0 · ~78GBRuns well
  • NVIDIA H100 (80GB)
    NVIDIA · Datacenter GPUs

    Fits at Q4_K_M (~44GB) with ~26.4GB headroom — about 1 concurrent instance.

    Q4_K_M · ~44GBRuns well

See all compatible devices →

Run Qwen2.5 72B on a private AI Business OS

Run your own AI agents on hardware you control — private by design, no per-seat data leaving your premises. BrainOutput helps you pick the right machine and turn it into a working AI Business OS.

Explore the AI Business OS