Benchmark result
mradermacher-GRM-3.2-Sky.i1-Q4_K_M on Intel Graphics — 14.7 tok/s
Measured with ik_llama.cpp:cpu manual on October 9, 2026.
What the model built
Open full screen →The coding scenario asks for a playable game in a single HTML file. This is exactly what the model returned, unedited; the frame only adds a line that tells this page how tall it is.
How this run compares
1st fastest of 4 runs of this model on Intel Graphics · median 14.7 tok/s.
sort
- 114.7tok/s77.1ik_llama.cpp:cpu · Q4_K_Mthis run
- 214.7tok/s81.6ik_llama.cpp:cpu · Q4_K_M
- 314.7tok/s81.7ik_llama.cpp:cpu · Q4_K_M
- 414.5tok/s81.2ik_llama.cpp:cpu · Q4_K_M
Reproduce this run
toolik_llama.cpp:cpu manual
modelmradermacher-GRM-3.2-Sky.i1-Q4_K_M (Q4_K_M)
context262,144
thinkingon
clientv0.4.87+102
Set those in ik_llama.cpp:cpu, then:
llm-benchmark benchmark --model "mradermacher-GRM-3.2-Sky.i1-Q4_K_M" --tool "ik_llama.cpp:cpu"The client prompts for context length and thinking mode, and for the KV cache dtype on Unsloth Studio. Sampling is left at the model default — the values above are what the tool reported using, not overrides the client sent.