llama.cpp on the Intel Arc B580 12GB

On the Intel Arc B580 12GB, llama.cpp runs GLM-4 9B at 4-bit at a median 21.3 tokens per second for generation, from 1 plain run (2K context).

Everything reported on the Intel Arc B580 12GB llama.cpp on other cards What fits on the Intel Arc B580 12GB

Speed by model on the Intel Arc B580 12GB

Plain runs only: one device, one request, no speculative decoding, the whole model in memory. Each row is one model size at one bit class, with the other engines' plain runs of the same model on this card beside it. Context length and build still differ between runs, and each side shows its own.

Model	llama.cpp	Other engines
GLM-4 9B 4-bit	21.3 t/sone report, 2K context	no other engine reported
Qwen3 4B 4-bit	12.7 t/sone report, context not stated	no other engine reported

4 llama.cpp reports on the Intel Arc B580 12GB

Newest first. A run on several cards, with speculative decoding, with part of the model in system RAM or with batched requests says so, since each describes a different setup.

Filter these reports on the homepage

Frequently asked

How fast is llama.cpp on the Intel Arc B580 12GB?

On the Intel Arc B580 12GB, llama.cpp runs GLM-4 9B at 4-bit at a median 21.3 tokens per second for generation, from 1 plain run (2K context).