RigRoute LabsRigRoute ↗
RigRoute Labs

Independent, reproducible hardware measurement for real hardware decisions.

Labs runs a fixed, versioned benchmark methodology on real hardware, using established tools (llama.cpp / llama-bench) rather than a homemade measurement engine, and publishes the exact command, raw output, and hardware/software evidence behind every number.

Current methodology

standard-ai-v0.1

RigRoute Labs' first local-LLM-inference methodology. Measures prompt-processing and generation throughput at a fixed context depth, across an ~8B and an ~14B Q4-family model, using llama-bench's own repetition/statistics support. llama.cpp build 11146 (commit 7fe450e19), context 8192, 5 repetitions per measurement.

Read the full methodology →

Latest tested hardware

Apple M1 Pro (32 GB unified memory)

ModelPrompt processingGeneration
llama 8B Q4_0190.4 tok/s27.9 tok/s
qwen3 14B Q4_K - Medium97.1 tok/s11.7 tok/s

AMD Radeon RX 9070 (16 GB dedicated VRAM)

ModelPrompt processingGeneration
llama 8B Q4_02511.8 tok/s102.3 tok/s
qwen3 14B Q4_K - Medium1177.2 tok/s56.4 tok/s
Benchmarks →
Every accepted measurement, in full.
Compare →
Side-by-side, no single score.
Data →
Raw, sanitized evidence for every run.