llamaperf

GLM-4

1 report

GLM-4 9B at 4-bit generates a median of 21.3 tokens per second on the Intel Arc B580 12GB, the hardware it is most run on, from one plain run (community reports, as of 22 Nov 2026).

As of 22 Nov 2026, GLM-4 9B at 4-bit on each device, fastest first, with the median generation speed of plain runs (one device, one request, no speculative decoding, the whole model in its memory):

Bars are medians on one scale, 0 to 21.3 tokens/s. Press ▶ to watch text arrive at a row's speed.

GLM-4 VRAM requirements by size and quant →

How does GLM-4 run on your hardware?

Pick your GPU or Mac, then read the speeds people reported for GLM-4 on it, or estimate memory fit and speed.

Free to use. No account needed. Memory estimates and community measurements are labelled separately.

Run GLM-4 yourself? Add your numbers.

Your coding agent can run a speed test on your machine and send it here with a free account. Or paste a result you've already posted.

Reported performance for GLM-4

Filter this model’s reports by setup →
One report for this model so far, so treat the figures as early. Add yours.

Speeds are as each source reported them. GPU count, offloading and concurrent requests change them, and prompt-processing speed is input throughput, not time to first token. Check the full setup before comparing two reports.

GLM-4 9B

Intel Arc B580 12GB · llama.cpp · 2,048 ctx

Tone: negativeIntel hardware
generation:
21.3 tokens/s
quant:
Q4_0 (GGUF)
kv:
F16

User reports GLM-4 9B at 21.25 t/s on an Intel Arc B580 12GB, with GPU utilization near 40%. Setup is llama.cpp (ipex-llm[cpp] build 2024.12.17) with Q4_0 GGUF and F16 KV cache, 2048 context, all 41 layers offloaded to the GPU. The user considers the performance low for the hardware. The prompt eval figure of 41.47 t/s is over only 5 tokens and is not reported as a prefill rate.

Oct 6, 2026

Get a weekly email of new GLM-4 reports on any GPU.

Email me new reports

Reported configurations

Individual observations from this page, not expected speeds or a ranking. Select a model to open its report, with the source and offloading conditions. Reported context may be a limit rather than actual input length.

Model / hardwareQuant / engineReported contextGeneration
GLM-4 9B
Intel Arc B580 12GB
Q4_0
llama.cpp
2,04821.3 tokens/s

Frequently asked

What hardware do people run GLM-4 on?

The hardware GLM-4 9B at 4-bit is most run on: Intel Arc B580 12GB (21.3 t/s median, one run). Medians of plain runs, as of 22 Nov 2026.

Which inference engines do people use for GLM-4?

Reports of GLM-4 name llama.cpp (1 report), as of 22 Nov 2026.