TECH RADAR / SERVING ENGINES

llama.cpp

Portable C/C++ inference across CPU and GPU.

Reviewed 16 Nov 2026Serving engines

What it covers

Local inference, quantized models, and CPU/GPU offload.

What we would check

Record the quantization format, backend, offloaded layers, and context length when measuring a run.

Go to the source

A curated ecosystem entry. The evaluation guidance is our interpretation, not a report of a completed test or a performance ranking.

← Back to Tech Radar