Boards / Raspberry Pi 5 cluster

Pi 5 cluster (distributed-llama): LLM benchmarks

Several Pi 5 boards splitting one model over Ethernet with distributed-llama.

At a glance

  • Measured runs2
  • Models tested1
  • Fastest run13.68 tok/s
  • Llama 3.2 3B Instruct, Q40, distributed-llama v0.12.3 · source ↗
  • Sources1

Where to buy

Every run on this board

ModelQuantRuntimeGen tok/sPrompt tok/sTTFT msSource
Llama 3.2 3B InstructQ40distributed-llama v0.12.313.6827.74–source ↗
2025-02-15
Llama 3.2 3B InstructQ40distributed-llama v0.12.310.3323.63–source ↗
2025-02-15

Can the Raspberry Pi 5 cluster run…

Can it run Llama 3.2 3B?

Have this board? Submit your own run →

Get the monthly board report

New boards, new runtimes, new numbers. One email a month, no hype.